1_1_10_Loan_Eligibility_Prediction_.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "source": [
        "# **<font color=\"blue\">Loan Eligibility Prediction Project Overview</font>**\n",
        "\n",
        "## **<font color=\"green\">1. Introduction</font>**\n",
        "\n",
        "Loan eligibility prediction is a critical task for financial institutions, helping them assess whether an applicant is likely to default on a loan or not. This project involves building a predictive model using machine learning techniques to determine loan eligibility based on applicant information.\n",
        "\n",
        "### **<font color=\"purple\">1.1 Importance of Loan Eligibility Prediction</font>**\n",
        "\n",
        "Predicting loan eligibility accurately is crucial for several reasons:\n",
        "- **Risk Management**: Financial institutions need to minimize the risk of loan defaults.\n",
        "- **Customer Satisfaction**: By quickly approving eligible loans, banks can improve customer satisfaction.\n",
        "- **Operational Efficiency**: Automating the eligibility process reduces manual effort and errors.\n",
        "\n",
        "### **<font color=\"purple\">1.2 Project Goals</font>**\n",
        "\n",
        "- To understand the data and perform exploratory data analysis (EDA).\n",
        "- To preprocess the data for machine learning.\n",
        "- To build and evaluate a Gradient Boosting Machine (GBM) model.\n",
        "- To make predictions on the test dataset.\n",
        "- To visualize model performance using various plots and metrics.\n",
        "\n",
        "### **<font color=\"purple\">1.3 Machine Learning Workflow</font>**\n",
        "\n",
        "The typical machine learning workflow involves several steps:\n",
        "1. **Data Collection**: Gathering the necessary data.\n",
        "2. **Data Cleaning and Preprocessing**: Handling missing values, encoding categorical variables, etc.\n",
        "3. **Exploratory Data Analysis (EDA)**: Understanding the data distribution and relationships.\n",
        "4. **Model Building**: Selecting and training a machine learning model.\n",
        "5. **Model Evaluation**: Assessing the model's performance using various metrics.\n",
        "6. **Model Deployment**: Implementing the model to make real-world predictions.\n",
        "\n",
        "## **<font color=\"green\">2. Dataset Overview</font>**\n",
        "\n",
        "We have two datasets: training data and test data.\n",
        "\n",
        "### **<font color=\"purple\">2.1 Training Data</font>**\n",
        "\n",
        "The training dataset contains information about loan applicants and whether they were granted a loan or not. Key features include:\n",
        "- **Applicant Income**: Income of the loan applicant.\n",
        "- **Coapplicant Income**: Income of the coapplicant.\n",
        "- **Loan Amount**: Requested loan amount.\n",
        "- **Loan Amount Term**: Term of the loan in months.\n",
        "- **Credit History**: Credit history of the applicant.\n",
        "- **Gender, Married, Dependents, Education, Self_Employed, Property_Area**: Demographic and categorical information.\n",
        "\n",
        "### **<font color=\"purple\">2.2 Test Data</font>**\n",
        "\n",
        "The test dataset contains similar information but without the loan status, which we aim to predict.\n",
        "\n",
        "## **<font color=\"green\">3. Data Loading and Initial Inspection</font>**\n",
        "\n",
        "### **<font color=\"purple\">3.1 Loading the Data</font>**\n",
        "\n",
        "Data loading is the first step in any data analysis project. We use the pandas library to load the datasets from CSV files.\n",
        "\n",
        "### **<font color=\"purple\">3.2 Inspecting the Data</font>**\n",
        "\n",
        "Initial inspection involves:\n",
        "- Displaying the first few rows of the dataset.\n",
        "- Checking the data types and missing values.\n",
        "- Displaying summary statistics to understand the distribution of numerical features.\n",
        "\n",
        "### **<font color=\"purple\">3.3 Understanding Data Types</font>**\n",
        "\n",
        "Understanding data types is essential for appropriate preprocessing. For instance, categorical variables need to be encoded into numerical format for machine learning algorithms to process them.\n",
        "\n",
        "## **<font color=\"green\">4. Exploratory Data Analysis (EDA)</font>**\n",
        "\n",
        "EDA is a crucial step to understand the data's underlying patterns, distributions, and relationships.\n",
        "\n",
        "### **<font color=\"purple\">4.1 Visualizing Income Distributions</font>**\n",
        "\n",
        "Histograms of applicant and coapplicant income provide insights into the distribution and central tendencies of income levels.\n",
        "\n",
        "### **<font color=\"purple\">4.2 Analyzing Loan Status Distribution</font>**\n",
        "\n",
        "A bar plot of loan status helps us understand the proportion of approved versus rejected loans.\n",
        "\n",
        "### **<font color=\"purple\">4.3 Examining Loan Amount by Education Level</font>**\n",
        "\n",
        "A box plot of loan amount by education level reveals how loan amounts vary with education, highlighting any disparities.\n",
        "\n",
        "### **<font color=\"purple\">4.4 Identifying Relationships and Patterns</font>**\n",
        "\n",
        "EDA helps identify relationships between features and the target variable (loan status), guiding the feature engineering and model selection process.\n",
        "\n",
        "## **<font color=\"green\">5. Data Preprocessing</font>**\n",
        "\n",
        "Data preprocessing prepares the data for machine learning by handling missing values, encoding categorical variables, and ensuring consistency across datasets.\n",
        "\n",
        "### **<font color=\"purple\">5.1 Handling Missing Values</font>**\n",
        "\n",
        "Missing values can skew the analysis and affect model performance. We handle missing values by:\n",
        "- Filling categorical columns with the most frequent value (mode).\n",
        "- Filling numerical columns with the median value to maintain distribution integrity.\n",
        "\n",
        "### **<font color=\"purple\">5.2 Encoding Categorical Variables</font>**\n",
        "\n",
        "Categorical variables are encoded using one-hot encoding to convert them into a numerical format suitable for machine learning algorithms.\n",
        "\n",
        "### **<font color=\"purple\">5.3 Ensuring Consistency Across Datasets</font>**\n",
        "\n",
        "After encoding, we ensure that both training and test datasets have the same columns, adding any missing columns to the test set with default values.\n",
        "\n",
        "### **<font color=\"purple\">5.4 Splitting the Data</font>**\n",
        "\n",
        "Splitting the data into training and validation sets allows us to evaluate model performance on unseen data, ensuring that the model generalizes well.\n",
        "\n",
        "## **<font color=\"green\">6. Model Building and Evaluation</font>**\n",
        "\n",
        "### **<font color=\"purple\">6.1 Choosing the Model</font>**\n",
        "\n",
        "We use the Gradient Boosting Machine (GBM) for its robustness and ability to handle various types of data effectively. GBM is an ensemble technique that combines the predictions of multiple base models to improve performance.\n",
        "\n",
        "### **<font color=\"purple\">6.2 Training the Model</font>**\n",
        "\n",
        "The model is trained on the training set, learning patterns and relationships within the data to make accurate predictions.\n",
        "\n",
        "### **<font color=\"purple\">6.3 Evaluating the Model</font>**\n",
        "\n",
        "Model evaluation involves:\n",
        "- **Accuracy**: Proportion of correct predictions.\n",
        "- **Precision**: Proportion of true positive predictions out of all positive predictions.\n",
        "- **Recall**: Proportion of true positive predictions out of all actual positives.\n",
        "- **F1 Score**: Harmonic mean of precision and recall.\n",
        "- **Confusion Matrix**: Detailed breakdown of true positives, true negatives, false positives, and false negatives.\n",
        "\n",
        "### **<font color=\"purple\">6.4 Receiver Operating Characteristic (ROC) Curve and AUC Score</font>**\n",
        "\n",
        "The ROC curve plots the true positive rate against the false positive rate, providing a visual representation of the model's performance across different thresholds. The AUC score quantifies the overall ability of the model to discriminate between positive and negative classes.\n",
        "\n",
        "## **<font color=\"green\">7. Making Predictions on Test Data</font>**\n",
        "\n",
        "### **<font color=\"purple\">7.1 Preparing the Test Data</font>**\n",
        "\n",
        "The test data is prepared by ensuring it has the same columns as the training data and is preprocessed similarly.\n",
        "\n",
        "### **<font color=\"purple\">7.2 Making Predictions</font>**\n",
        "\n",
        "The trained model is used to make predictions on the test dataset, determining the loan eligibility status for each applicant.\n",
        "\n",
        "### **<font color=\"purple\">7.3 Preparing the Submission File</font>**\n",
        "\n",
        "The predictions are formatted into a submission file, typically required in competition or real-world scenarios, detailing the loan status for each applicant in the test set.\n",
        "\n",
        "## **<font color=\"green\">8. Conclusion</font>**\n",
        "\n",
        "### **<font color=\"purple\">8.1 Summary of Achievements</font>**\n",
        "\n",
        "- We successfully built a predictive model for loan eligibility.\n",
        "- Performed comprehensive EDA to understand the data.\n",
        "- Preprocessed the data to handle missing values and encode categorical variables.\n",
        "- Trained and evaluated a GBM model, achieving high performance on validation data.\n",
        "- Made predictions on test data and prepared a submission file.\n",
        "\n",
        "### **<font color=\"purple\">8.2 Importance of Each Step</font>**\n",
        "\n",
        "- **EDA**: Helped in understanding data distribution and identifying key patterns.\n",
        "- **Preprocessing**: Ensured data was clean and in the right format for machine learning.\n",
        "- **Model Training and Evaluation**: Enabled us to build a robust model and assess its performance.\n",
        "- **Prediction and Submission**: Demonstrated the practical application of the model.\n",
        "\n",
        "### **<font color=\"purple\">8.3 Further Steps</font>**\n",
        "\n",
        "- **Experiment with Different Models**: Trying other machine learning models like Random Forest, XGBoost, etc., to compare performance.\n",
        "- **Hyperparameter Tuning**: Fine-tuning model parameters to improve accuracy.\n",
        "- **Feature Engineering**: Creating new features based on domain knowledge to enhance model performance.\n",
        "- **Cross-Validation**: Using cross-validation techniques to ensure the model's robustness.\n",
        "\n",
        "## **<font color=\"green\">9. Additional Learning Resources</font>**\n",
        "\n",
        "### **<font color=\"purple\">9.1 Books and Articles</font>**\n",
        "\n",
        "- **\"Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow\" by Aurélien Géron**: A comprehensive guide to machine learning.\n",
        "- **\"Pattern Recognition and Machine Learning\" by Christopher Bishop**: An in-depth book on machine learning theories and practices.\n",
        "- **Machine Learning Articles on Medium**: Regularly updated articles on various machine learning topics.\n",
        "\n",
        "### **<font color=\"purple\">9.2 Online Courses</font>**\n",
        "\n",
        "- **Coursera: Machine Learning by Andrew Ng**: A popular and highly recommended course.\n",
        "- **Udemy: Complete Machine Learning and Data Science Bootcamp**: Comprehensive course covering all aspects of machine learning.\n",
        "- **edX: MicroMasters in Artificial Intelligence**: Series of graduate-level courses in AI and machine learning.\n",
        "\n",
        "### **<font color=\"purple\">9.3 Practice Platforms</font>**\n",
        "\n",
        "- **Kaggle**: Provides datasets, competitions, and community support for machine learning projects.\n",
        "- **DataCamp**: Interactive learning platform for data science and machine learning.\n",
        "\n",
        "---\n",
        "\n",
        "By following this detailed overview and utilizing the additional resources, students will gain a solid understanding of the machine learning process and be well-equipped to tackle similar projects independently.\n",
        "\n",
        "---\n",
        "\n",
        "**End of Project Overview**"
      ],
      "metadata": {
        "id": "I8CyQC40rdqL"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# **<font color=\"blue\">Importing Essential Libraries for Machine Learning</font>**\n",
        "\n",
        "## **<font color=\"green\">1. Introduction</font>**\n",
        "\n",
        "In any machine learning project, importing the right libraries is crucial for data manipulation, visualization, model building, and evaluation. This section explores the essential libraries used in the Python ecosystem for machine learning and data analysis.\n",
        "\n",
        "### **<font color=\"purple\">1.1 Importance of Libraries in Machine Learning</font>**\n",
        "\n",
        "Libraries provide pre-built functionalities and tools that significantly expedite the development and deployment of machine learning models. They offer:\n",
        "- **Efficiency**: Ready-made functions for common tasks like data cleaning, preprocessing, and model evaluation.\n",
        "- **Flexibility**: Support for various algorithms, allowing experimentation and model comparison.\n",
        "- **Visualization**: Tools to create informative plots and graphs for data exploration and model performance.\n",
        "\n",
        "## **<font color=\"green\">2. Libraries Overview</font>**\n",
        "\n",
        "### **<font color=\"purple\">2.1 Pandas (pd)</font>**\n",
        "\n",
        "- **Functionality**: Pandas is used for data manipulation and analysis.\n",
        "- **Key Features**: DataFrames for structured data, powerful methods for filtering, grouping, and joining datasets.\n",
        "- **Usage**: Loading datasets from files (CSV, Excel), handling missing data, transforming data for modeling.\n",
        "\n",
        "### **<font color=\"purple\">2.2 Matplotlib.pyplot (plt)</font>**\n",
        "\n",
        "- **Functionality**: Matplotlib is a plotting library for creating static, animated, and interactive visualizations.\n",
        "- **Key Features**: Support for various plot types (line plots, scatter plots, histograms), customization of plot aesthetics.\n",
        "- **Usage**: Visualizing data distributions, trends, correlations, and model performance metrics like ROC curves and confusion matrices.\n",
        "\n",
        "### **<font color=\"purple\">2.3 Seaborn (sns)</font>**\n",
        "\n",
        "- **Functionality**: Seaborn is built on top of Matplotlib and provides a high-level interface for drawing attractive statistical graphics.\n",
        "- **Key Features**: Simplified syntax, default themes, support for complex plots like heatmaps, violin plots, and categorical plots.\n",
        "- **Usage**: Enhancing the visual appeal of plots, exploring relationships between variables in datasets.\n",
        "\n",
        "### **<font color=\"purple\">2.4 Scikit-Learn (sklearn)</font>**\n",
        "\n",
        "- **Functionality**: Scikit-Learn is a machine learning library providing simple and efficient tools for data mining and analysis.\n",
        "- **Key Features**: Implementation of various machine learning algorithms (classification, regression, clustering), model evaluation metrics.\n",
        "- **Usage**: Building, training, and evaluating machine learning models, handling preprocessing tasks like scaling and encoding.\n",
        "\n",
        "### **<font color=\"purple\">2.5 Additional Libraries</font>**\n",
        "\n",
        "- **NumPy (np)**: Fundamental package for scientific computing in Python, providing support for large, multi-dimensional arrays and matrices.\n",
        "- **Statsmodels**: Library for estimating and testing statistical models.\n",
        "- **TensorFlow, PyTorch**: Deep learning libraries for building and training neural networks.\n",
        "- **NLTK, SpaCy**: Libraries for natural language processing tasks like tokenization, part-of-speech tagging, and named entity recognition.\n",
        "\n",
        "## **<font color=\"green\">3. Installation and Environment Setup</font>**\n",
        "\n",
        "Ensure that all required libraries are installed in your Python environment using pip or conda:\n",
        "```bash\n",
        "pip install pandas matplotlib seaborn scikit-learn numpy"
      ],
      "metadata": {
        "id": "N35GuhyosxiF"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import pandas as pd  # Importing pandas for data manipulation and analysis"
      ],
      "metadata": {
        "id": "VwJ6RR9EkTP_"
      },
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt  # Importing matplotlib for plotting graphs"
      ],
      "metadata": {
        "id": "3khMf13jlKMw"
      },
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import seaborn as sns  # Importing seaborn for enhanced plotting capabilities"
      ],
      "metadata": {
        "id": "ka5wAw1XlKPx"
      },
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.model_selection import train_test_split  # Importing train_test_split for splitting data into training and validation sets\n"
      ],
      "metadata": {
        "id": "6qHWCiKRlKSB"
      },
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.ensemble import GradientBoostingClassifier  # Importing GradientBoostingClassifier for training the model\n"
      ],
      "metadata": {
        "id": "fvm5ASj4lKWa"
      },
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score, confusion_matrix, roc_curve, roc_auc_score  # Importing metrics for model evaluation\n"
      ],
      "metadata": {
        "id": "b71lGh0mlKaB"
      },
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "xyj2Aokzk1ef"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "NWa5_RmCs4iU"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# **<font color=\"blue\">Step 1: Data Collection in Machine Learning Projects</font>**\n",
        "\n",
        "## **<font color=\"green\">1. Introduction</font>**\n",
        "\n",
        "Data collection is the foundational step in any machine learning project. It involves gathering relevant datasets that will be used for training, validating, and testing machine learning models. This phase is crucial as the quality and quantity of data directly impact the model's performance and reliability.\n",
        "\n",
        "### **<font color=\"purple\">1.1 Importance of Data Collection</font>**\n",
        "\n",
        "- **Foundation of ML Models**: High-quality data ensures robust and accurate machine learning models.\n",
        "- **Bias Mitigation**: Proper data collection helps mitigate biases and ensures fairness in predictions.\n",
        "- **Decision Making**: Data-driven decisions are crucial for business insights and strategy formulation.\n",
        "- **Model Performance**: Data quantity and diversity improve the generalizability and performance of models.\n",
        "\n",
        "## **<font color=\"green\">2. Sources of Data</font>**\n",
        "\n",
        "### **<font color=\"purple\">2.1 Public Datasets</font>**\n",
        "\n",
        "- **Advantages**: Free access, diverse domains, already cleaned and formatted.\n",
        "- **Examples**: UCI Machine Learning Repository, Kaggle Datasets, government data portals.\n",
        "\n",
        "### **<font color=\"purple\">2.2 Private Datasets</font>**\n",
        "\n",
        "- **Advantages**: Specific to business needs, proprietary information.\n",
        "- **Considerations**: Legal and privacy concerns, data governance, consent.\n",
        "\n",
        "### **<font color=\"purple\">2.3 Data Scraping</font>**\n",
        "\n",
        "- **Purpose**: Extract data from websites or APIs.\n",
        "- **Tools**: BeautifulSoup, Scrapy for web scraping; Requests for API integration.\n",
        "- **Ethical Considerations**: Respect robots.txt, terms of service, and copyright laws.\n",
        "\n",
        "### **<font color=\"purple\">2.4 Data Generation</font>**\n",
        "\n",
        "- **Purpose**: Simulate data for specific scenarios or augment existing datasets.\n",
        "- **Methods**: Synthetic data generation, data augmentation techniques.\n",
        "- **Use Cases**: Imbalanced class handling, privacy-preserving techniques.\n",
        "\n",
        "## **<font color=\"green\">3. Methods of Data Collection</font>**\n",
        "\n",
        "### **<font color=\"purple\">3.1 Surveys and Questionnaires</font>**\n",
        "\n",
        "- **Purpose**: Collect subjective data from individuals or organizations.\n",
        "- **Challenges**: Response bias, data validation, sample size.\n",
        "\n",
        "### **<font color=\"purple\">3.2 Observational Studies</font>**\n",
        "\n",
        "- **Purpose**: Passive collection of data without intervention.\n",
        "- **Examples**: Tracking user behavior, environmental data collection.\n",
        "\n",
        "### **<font color=\"purple\">3.3 Experimental Design</font>**\n",
        "\n",
        "- **Purpose**: Controlled environment to test hypotheses.\n",
        "- **Examples**: A/B testing, clinical trials.\n",
        "\n",
        "### **<font color=\"purple\">3.4 Data Integration and Fusion</font>**\n",
        "\n",
        "- **Purpose**: Combine data from multiple sources to enhance insights.\n",
        "- **Challenges**: Data compatibility, schema matching, data deduplication.\n",
        "\n",
        "## **<font color=\"green\">4. Data Collection Best Practices</font>**\n",
        "\n",
        "### **<font color=\"purple\">4.1 Define Objectives and Requirements</font>**\n",
        "\n",
        "- **Clarify**: Purpose of data collection, specific variables of interest, sample size requirements.\n",
        "\n",
        "### **<font color=\"purple\">4.2 Data Quality Assurance</font>**\n",
        "\n",
        "- **Ensure**: Data accuracy, completeness, consistency, and relevance.\n",
        "- **Tools**: Data profiling, outlier detection, missing value handling.\n",
        "\n",
        "### **<font color=\"purple\">4.3 Ethical Considerations</font>**\n",
        "\n",
        "- **Privacy**: Ensure data anonymization, consent, and compliance with GDPR or local regulations.\n",
        "- **Bias**: Mitigate biases in data collection and usage to ensure fairness.\n",
        "\n",
        "### **<font color=\"purple\">4.4 Data Documentation and Metadata</font>**\n",
        "\n",
        "- **Document**: Data sources, collection methods, preprocessing steps.\n",
        "- **Metadata**: Schema information, data lineage, version control.\n",
        "\n",
        "## **<font color=\"green\">5. Conclusion</font>**\n",
        "\n",
        "Data collection is a critical phase in machine learning projects, influencing the success and reliability of models. By following best practices, addressing ethical considerations, and leveraging diverse data sources, practitioners can ensure robust data foundations for their machine learning endeavors.\n",
        "\n",
        "---\n",
        "\n",
        "These comprehensive notes cover various aspects of data collection in machine learning projects, emphasizing its importance, methods, sources, challenges, and best practices. They provide a detailed framework for effectively gathering and managing data for successful machine learning model development."
      ],
      "metadata": {
        "id": "8eZDZk2As44c"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Load the training and test datasets"
      ],
      "metadata": {
        "id": "Kq-IXF3_k1hn"
      },
      "execution_count": 77,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "train_data = pd.read_csv('train_u6lujuX_CVtuZ9i.csv')  # Loading training data from CSV file"
      ],
      "metadata": {
        "id": "eP5kSnnfk1kP"
      },
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "test_data = pd.read_csv('test_Y3wMUE5_7gLdaTN.csv')  # Loading test data from CSV file"
      ],
      "metadata": {
        "id": "aNs0W4cqk1oy"
      },
      "execution_count": 10,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "Ps5pkxFzk1rR"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "eIloH0D8tIkg"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# <font color='blue'>Step 2: Displaying Basic Information about the Training and Test Data</font>\n",
        "\n",
        "In Step 2, the objective is to gain a comprehensive understanding of the structure and essential characteristics of both the training and test datasets. This foundational step is crucial as it provides insights into the dataset's format, features, and initial data quality assessment.\n",
        "\n",
        "## <font color='purple'>Importance of Understanding Data Structure</font>\n",
        "\n",
        "Understanding the structure of the data is pivotal for effective data analysis and modeling. It involves checking the initial rows of the dataset, examining data types, and summarizing key statistics to ensure data integrity and inform subsequent steps in the data science pipeline.\n",
        "\n",
        "### <font color='green'>Techniques Used in Step 2</font>\n",
        "\n",
        "#### Loading the Datasets\n",
        "\n",
        "- **pandas Library**: Use `pandas` to load CSV files (`read_csv` function) containing the training and test data. This step involves importing necessary libraries and reading data files into pandas DataFrames.\n",
        "\n",
        "#### Displaying Data Characteristics\n",
        "\n",
        "- **`head()` Function**: Display the first few rows of each dataset (`train_data.head()` and `test_data.head()`) to obtain a preview of the data structure, feature names, and initial values.\n",
        "\n",
        "#### Understanding Dataset Information\n",
        "\n",
        "- **`info()` Function**: Retrieve a concise summary of each dataset (`train_data.info()` and `test_data.info()`), which includes the number of entries, data types, and memory usage. This function provides an overview of the dataset's composition and identifies any missing or inconsistent data types.\n",
        "\n",
        "#### Summary Statistics\n",
        "\n",
        "- **`describe()` Function**: Generate descriptive statistics (`train_data.describe()` and `test_data.describe()`) that summarize the central tendency, dispersion, and distribution shape of numerical features. This step helps in identifying outliers and understanding the overall data distribution.\n",
        "\n",
        "## <font color='orange'>Conclusion</font>\n",
        "\n",
        "Step 2 lays the foundation for data exploration and preprocessing in the machine learning workflow. By understanding the dataset's structure and characteristics early on, data scientists can make informed decisions regarding data cleaning, feature engineering, and model selection.\n",
        "\n",
        "<div style=\"background-color: #FFFACD; padding: 10px; border-radius: 5px;\">\n",
        "    <p style=\"color: #FF6347; font-weight: bold;\">Tip:</p>\n",
        "    <p style=\"color: #FF6347;\">Visualize data distributions and identify potential outliers or missing values during the initial data inspection.</p>\n",
        "</div>\n",
        "\n",
        "<div style=\"background-color: #F0E68C; padding: 10px; border-radius: 5px;\">\n",
        "    <p style=\"color: #B8860B; font-weight: bold;\">Note:</p>\n",
        "    <p style=\"color: #B8860B;\">Ensure consistency in data loading and format understanding across both training and test datasets.</p>\n",
        "</div>\n",
        "\n",
        "---\n",
        "\n",
        "By following Step 2 diligently, data scientists can establish a solid understanding of the dataset's initial characteristics and ensure readiness for subsequent data preprocessing and exploratory analysis tasks. This approach enhances the efficiency and effectiveness of the overall data science workflow."
      ],
      "metadata": {
        "id": "Ph59rdHLyRB_"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Display the first few rows of the training data"
      ],
      "metadata": {
        "id": "gCySV-B6k1tB"
      },
      "execution_count": 78,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(\"Training Data:\")  # Printing a header for training data\n",
        "print(train_data.head())  # Displaying the first few rows of the training data"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "W0nCO-Uok1vL",
        "outputId": "54ef9722-f9d8-4060-864c-bc5181a0c351"
      },
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Training Data:\n",
            "    Loan_ID Gender Married Dependents     Education Self_Employed  \\\n",
            "0  LP001002   Male      No          0      Graduate            No   \n",
            "1  LP001003   Male     Yes          1      Graduate            No   \n",
            "2  LP001005   Male     Yes          0      Graduate           Yes   \n",
            "3  LP001006   Male     Yes          0  Not Graduate            No   \n",
            "4  LP001008   Male      No          0      Graduate            No   \n",
            "\n",
            "   ApplicantIncome  CoapplicantIncome  LoanAmount  Loan_Amount_Term  \\\n",
            "0             5849                0.0         NaN             360.0   \n",
            "1             4583             1508.0       128.0             360.0   \n",
            "2             3000                0.0        66.0             360.0   \n",
            "3             2583             2358.0       120.0             360.0   \n",
            "4             6000                0.0       141.0             360.0   \n",
            "\n",
            "   Credit_History Property_Area Loan_Status  \n",
            "0             1.0         Urban           Y  \n",
            "1             1.0         Rural           N  \n",
            "2             1.0         Urban           Y  \n",
            "3             1.0         Urban           Y  \n",
            "4             1.0         Urban           Y  \n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "Psxxc5iIk1xL"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "FcocPC0nk105"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Display basic information about the training data"
      ],
      "metadata": {
        "id": "lcei9r1ak13h"
      },
      "execution_count": 79,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(\"\\nTraining Data Info:\")  # Printing a header for training data info\n",
        "print(train_data.info())  # Displaying info about the training data"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "pocHWDNhk17x",
        "outputId": "90587956-ea1d-462b-db5e-4761d79f93c7"
      },
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "Training Data Info:\n",
            "<class 'pandas.core.frame.DataFrame'>\n",
            "RangeIndex: 614 entries, 0 to 613\n",
            "Data columns (total 13 columns):\n",
            " #   Column             Non-Null Count  Dtype  \n",
            "---  ------             --------------  -----  \n",
            " 0   Loan_ID            614 non-null    object \n",
            " 1   Gender             601 non-null    object \n",
            " 2   Married            611 non-null    object \n",
            " 3   Dependents         599 non-null    object \n",
            " 4   Education          614 non-null    object \n",
            " 5   Self_Employed      582 non-null    object \n",
            " 6   ApplicantIncome    614 non-null    int64  \n",
            " 7   CoapplicantIncome  614 non-null    float64\n",
            " 8   LoanAmount         592 non-null    float64\n",
            " 9   Loan_Amount_Term   600 non-null    float64\n",
            " 10  Credit_History     564 non-null    float64\n",
            " 11  Property_Area      614 non-null    object \n",
            " 12  Loan_Status        614 non-null    object \n",
            "dtypes: float64(4), int64(1), object(8)\n",
            "memory usage: 62.5+ KB\n",
            "None\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "PBuvpS_tmTtg"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "HoOUARd-mTwt"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Display the first few rows of the test data"
      ],
      "metadata": {
        "id": "zSt8IWKBmTzW"
      },
      "execution_count": 80,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(\"\\nTest Data:\")  # Printing a header for test data\n",
        "print(test_data.head())  # Displaying the first few rows of the test data"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "BCmC1sYImT1x",
        "outputId": "897b6ffa-4adf-42c4-adea-a8ffc429d2b8"
      },
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "Test Data:\n",
            "    Loan_ID Gender Married Dependents     Education Self_Employed  \\\n",
            "0  LP001015   Male     Yes          0      Graduate            No   \n",
            "1  LP001022   Male     Yes          1      Graduate            No   \n",
            "2  LP001031   Male     Yes          2      Graduate            No   \n",
            "3  LP001035   Male     Yes          2      Graduate            No   \n",
            "4  LP001051   Male      No          0  Not Graduate            No   \n",
            "\n",
            "   ApplicantIncome  CoapplicantIncome  LoanAmount  Loan_Amount_Term  \\\n",
            "0             5720                  0       110.0             360.0   \n",
            "1             3076               1500       126.0             360.0   \n",
            "2             5000               1800       208.0             360.0   \n",
            "3             2340               2546       100.0             360.0   \n",
            "4             3276                  0        78.0             360.0   \n",
            "\n",
            "   Credit_History Property_Area  \n",
            "0             1.0         Urban  \n",
            "1             1.0         Urban  \n",
            "2             1.0         Urban  \n",
            "3             NaN         Urban  \n",
            "4             1.0         Urban  \n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "fUFFmL9Ek1-R"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "0YGl7A9Sk2AQ"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Display basic information about the test data"
      ],
      "metadata": {
        "id": "lX8kbmzlk2Cq"
      },
      "execution_count": 81,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(\"\\nTest Data Info:\")  # Printing a header for test data info\n",
        "print(test_data.info())  # Displaying info about the test data"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "gOSZ0aPXk2E2",
        "outputId": "544f598d-a8e4-4ab1-b88d-27b7e2b5711f"
      },
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "Test Data Info:\n",
            "<class 'pandas.core.frame.DataFrame'>\n",
            "RangeIndex: 367 entries, 0 to 366\n",
            "Data columns (total 12 columns):\n",
            " #   Column             Non-Null Count  Dtype  \n",
            "---  ------             --------------  -----  \n",
            " 0   Loan_ID            367 non-null    object \n",
            " 1   Gender             356 non-null    object \n",
            " 2   Married            367 non-null    object \n",
            " 3   Dependents         357 non-null    object \n",
            " 4   Education          367 non-null    object \n",
            " 5   Self_Employed      344 non-null    object \n",
            " 6   ApplicantIncome    367 non-null    int64  \n",
            " 7   CoapplicantIncome  367 non-null    int64  \n",
            " 8   LoanAmount         362 non-null    float64\n",
            " 9   Loan_Amount_Term   361 non-null    float64\n",
            " 10  Credit_History     338 non-null    float64\n",
            " 11  Property_Area      367 non-null    object \n",
            "dtypes: float64(3), int64(2), object(7)\n",
            "memory usage: 34.5+ KB\n",
            "None\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "7WpF2oIck2G3"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "jA9LELPlk2JJ"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Display summary statistics of the training data"
      ],
      "metadata": {
        "id": "SjOHT_oEk2Lh"
      },
      "execution_count": 82,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(\"\\nTraining Data Summary Statistics:\")  # Printing a header for summary statistics\n",
        "print(train_data.describe())  # Displaying summary statistics of the training data"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "GsJr5vd4kTTA",
        "outputId": "8d4bbffb-5484-46e1-b70f-6182d4c9c3a5"
      },
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "Training Data Summary Statistics:\n",
            "       ApplicantIncome  CoapplicantIncome  LoanAmount  Loan_Amount_Term  \\\n",
            "count       614.000000         614.000000  592.000000         600.00000   \n",
            "mean       5403.459283        1621.245798  146.412162         342.00000   \n",
            "std        6109.041673        2926.248369   85.587325          65.12041   \n",
            "min         150.000000           0.000000    9.000000          12.00000   \n",
            "25%        2877.500000           0.000000  100.000000         360.00000   \n",
            "50%        3812.500000        1188.500000  128.000000         360.00000   \n",
            "75%        5795.000000        2297.250000  168.000000         360.00000   \n",
            "max       81000.000000       41667.000000  700.000000         480.00000   \n",
            "\n",
            "       Credit_History  \n",
            "count      564.000000  \n",
            "mean         0.842199  \n",
            "std          0.364878  \n",
            "min          0.000000  \n",
            "25%          1.000000  \n",
            "50%          1.000000  \n",
            "75%          1.000000  \n",
            "max          1.000000  \n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "N4j2S7SHmu1Y"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "5NYYqb-4mu4N"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# <font color='blue'>Step 3: Exploratory Data Analysis (EDA)</font>\n",
        "\n",
        "## <font color='green'>Overview</font>\n",
        "Exploratory Data Analysis (EDA) is a critical phase in any data science or machine learning project. It involves examining and visualizing data to understand its characteristics, detect patterns, spot anomalies, and inform data preprocessing steps. EDA lays the groundwork for building accurate and reliable machine learning models.\n",
        "\n",
        "## <font color='green'>Importance of EDA:</font>\n",
        "\n",
        "- **Understanding Data Structure**: EDA provides insights into the dataset's dimensions, features, and data types, helping in initial data exploration.\n",
        "\n",
        "- **Detecting Data Quality Issues**: By analyzing summary statistics and visualizations, EDA identifies missing values, outliers, and inconsistencies that require cleaning.\n",
        "\n",
        "- **Informing Feature Engineering**: EDA guides feature selection and creation by revealing relationships and distributions within the data.\n",
        "\n",
        "- **Visualizing Relationships**: Techniques like scatter plots, histograms, and correlation matrices uncover dependencies and interactions between variables.\n",
        "\n",
        "## <font color='green'>Techniques Used in EDA:</font>\n",
        "\n",
        "### <font color='purple'>1. Summary Statistics:</font>\n",
        "- **Central Tendency**: Mean, median, mode\n",
        "- **Dispersion**: Standard deviation, range, quartiles\n",
        "\n",
        "### <font color='purple'>2. Univariate Analysis:</font>\n",
        "- **Histograms**: Explore distributions of numerical variables.\n",
        "- **Bar Plots**: Display frequencies of categorical variables.\n",
        "- **Box Plots**: Visualize data spread and identify outliers.\n",
        "\n",
        "### <font color='purple'>3. Bivariate Analysis:</font>\n",
        "- **Scatter Plots**: Investigate relationships between pairs of numerical variables.\n",
        "- **Pair Plots**: Grid of scatter plots for multiple numerical variables.\n",
        "- **Heatmaps**: Represent correlation matrices with color gradients.\n",
        "\n",
        "### <font color='purple'>4. Handling Missing Data:</font>\n",
        "- Identify missing values and choose appropriate imputation techniques based on data context.\n",
        "\n",
        "### <font color='purple'>5. Outlier Detection and Treatment:</font>\n",
        "- Use visualizations like box plots and scatter plots to detect outliers and decide on appropriate actions.\n",
        "\n",
        "### <font color='purple'>6. Feature Engineering:</font>\n",
        "- Create new features from existing ones to capture relevant information for model training.\n",
        "\n",
        "## <font color='green'>Example Visualizations:</font>\n",
        "\n",
        "- **Histogram of Applicant Income**:\n",
        "  - Illustrates income distribution among applicants, highlighting income levels and potential skewness.\n",
        "\n",
        "- **Histogram of Coapplicant Income**:\n",
        "  - Shows distribution of coapplicant incomes, indicating their role in loan applications.\n",
        "\n",
        "- **Bar Plot of Loan Status**:\n",
        "  - Visualizes loan approval outcomes ('Y' for Yes, 'N' for No), providing insights into dataset balance.\n",
        "\n",
        "- **Box Plot of Loan Amount by Education**:\n",
        "  - Compares loan amounts across different education levels to discern trends and variations.\n",
        "\n",
        "## <font color='green'>Conclusion:</font>\n",
        "EDA is an iterative process that continues throughout the machine learning pipeline. It facilitates data understanding, prepares data for modeling, and ensures the reliability of insights drawn. Effective EDA enhances model performance and helps in making informed decisions based on data exploration.\n",
        "\n",
        "<div style=\"background-color: #FFFACD; padding: 10px; border-radius: 5px;\">\n",
        "    <p style=\"color: #FF6347; font-weight: bold;\">Tip:</p>\n",
        "    <p style=\"color: #FF6347;\">Use interactive visualization tools like Plotly or Tableau for dynamic EDA.</p>\n",
        "</div>\n",
        "\n",
        "<div style=\"background-color: #F0E68C; padding: 10px; border-radius: 5px;\">\n",
        "    <p style=\"color: #B8860B; font-weight: bold;\">Note:</p>\n",
        "    <p style=\"color: #B8860B;\">EDA should be tailored to specific project goals and data characteristics.</p>\n",
        "</div>\n",
        "\n",
        "---\n",
        "\n",
        "### <font color='green'>Additional Considerations in EDA:</font>\n",
        "\n",
        "#### <font color='purple'>Exploring Data Distributions:</font>\n",
        "Understanding the distribution of each variable is crucial for feature selection and model assumptions. Visualizing histograms and density plots helps in identifying the shape, central tendency, and spread of data.\n",
        "\n",
        "#### <font color='purple'>Handling Categorical Data:</font>\n",
        "For categorical variables, bar plots and frequency tables are effective in visualizing class distributions. One-hot encoding or label encoding transforms categorical data into numerical formats suitable for modeling.\n",
        "\n",
        "#### <font color='purple'>Correlation Analysis:</font>\n",
        "Correlation matrices and heatmaps reveal relationships between numerical variables. Strong correlations may indicate multicollinearity, influencing feature selection and model performance.\n",
        "\n",
        "#### <font color='purple'>Time-Series Analysis:</font>\n",
        "In time-series data, line plots and lag plots help in identifying trends, seasonality, and autocorrelation. Understanding temporal patterns is critical for accurate forecasting and anomaly detection.\n",
        "\n",
        "#### <font color='purple'>Geospatial Analysis:</font>\n",
        "For spatial data, maps and choropleth plots visualize geographical patterns and distributions. Geospatial analysis provides insights into regional variations and spatial dependencies.\n",
        "\n",
        "---\n",
        "\n",
        "### <font color='green'>Best Practices for Effective EDA:</font>\n",
        "\n",
        "#### <font color='purple'>1. Iterative Approach:</font>\n",
        "EDA is not a one-time task; it evolves as insights are gained and new questions arise. Iteratively explore different aspects of data to uncover hidden patterns and refine hypotheses.\n",
        "\n",
        "#### <font color='purple'>2. Visual Storytelling:</font>\n",
        "Use clear and insightful visualizations to communicate findings effectively. Proper labeling, color choices, and annotations enhance the readability and interpretability of plots.\n",
        "\n",
        "#### <font color='purple'>3. Data Driven Decision Making:</font>\n",
        "Let data exploration guide decision-making processes. Validate assumptions, refine hypotheses, and adjust modeling strategies based on EDA insights.\n",
        "\n",
        "#### <font color='purple'>4. Documentation:</font>\n",
        "Document findings, observations, and data transformations during EDA. Clear documentation ensures reproducibility and facilitates collaboration across team members.\n",
        "\n",
        "---\n",
        "\n",
        "### <font color='green'>Conclusion:</font>\n",
        "\n",
        "Exploratory Data Analysis is a foundational step in the machine learning journey, enabling data scientists to understand data characteristics, detect patterns, and prepare data for modeling. By employing a variety of techniques and visualizations, EDA transforms raw data into actionable insights, driving informed decisions and robust model development.\n",
        "\n",
        "<div style=\"background-color: #ADD8E6; padding: 10px; border-radius: 5px;\">\n",
        "    <p style=\"color: #4682B4; font-weight: bold;\">Key Takeaway:</p>\n",
        "    <p style=\"color: #4682B4;\">Investing time in thorough EDA enhances model performance and ensures data-driven decision-making.</p>\n",
        "</div>"
      ],
      "metadata": {
        "id": "d-64IH3_tiBk"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#  Exploratory Data Analysis (EDA)"
      ],
      "metadata": {
        "id": "DKJRbPz-pdMh"
      },
      "execution_count": 83,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Histogram of Applicant Income"
      ],
      "metadata": {
        "id": "MjNV4MhCpdPp"
      },
      "execution_count": 62,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "plt.figure(figsize=(10, 6))  # Setting figure size\n",
        "plt.hist(train_data['ApplicantIncome'], bins=20, color='skyblue', edgecolor='black')  # Plotting histogram\n",
        "plt.title('Histogram of Applicant Income')  # Adding title\n",
        "plt.xlabel('Applicant Income')  # Adding x-axis label\n",
        "plt.ylabel('Frequency')  # Adding y-axis label\n",
        "plt.grid(True)  # Enabling grid\n",
        "plt.show()  # Displaying plot"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 564
        },
        "id": "JUix_pQIpi2y",
        "outputId": "042e6813-d93c-49d0-80f4-a97a26d4fa0c"
      },
      "execution_count": 63,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1000x600 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "-7oSfxWMpjAi"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "KBW0_75MpdTE"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Histogram of Coapplicant Income"
      ],
      "metadata": {
        "id": "SmHS4MmrpnLK"
      },
      "execution_count": 64,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "plt.figure(figsize=(10, 6))  # Setting figure size\n",
        "plt.hist(train_data['CoapplicantIncome'], bins=20, color='lightgreen', edgecolor='black')  # Plotting histogram\n",
        "plt.title('Histogram of Coapplicant Income')  # Adding title\n",
        "plt.xlabel('Coapplicant Income')  # Adding x-axis label\n",
        "plt.ylabel('Frequency')  # Adding y-axis label\n",
        "plt.grid(True)  # Enabling grid\n",
        "plt.show()  # Displaying plot"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 564
        },
        "id": "jLaOJkH4pnOV",
        "outputId": "d4378cf1-3b3d-43da-a3e3-cb0a5eb35550"
      },
      "execution_count": 65,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1000x600 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "7FB3rTx7pnRS"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "XnvN_ilypnUI"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Bar plot of Loan Status"
      ],
      "metadata": {
        "id": "jmOXt78ZpxKT"
      },
      "execution_count": 66,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "plt.figure(figsize=(8, 5))  # Setting figure size\n",
        "sns.countplot(train_data['Loan_Status'], palette='pastel')  # Plotting bar plot\n",
        "plt.title('Loan Status Distribution')  # Adding title\n",
        "plt.xlabel('Loan Status')  # Adding x-axis label\n",
        "plt.ylabel('Count')  # Adding y-axis label\n",
        "plt.show()  # Displaying plot"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 594
        },
        "id": "QlLYueHNpxNc",
        "outputId": "ff7924d7-435e-446a-b821-c13d32b6167d"
      },
      "execution_count": 67,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "<ipython-input-67-b67a62979908>:2: FutureWarning: \n",
            "\n",
            "Passing `palette` without assigning `hue` is deprecated and will be removed in v0.14.0. Assign the `y` variable to `hue` and set `legend=False` for the same effect.\n",
            "\n",
            "  sns.countplot(train_data['Loan_Status'], palette='pastel')  # Plotting bar plot\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 800x500 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "TtTL2sMqp4yi"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "K3NnvIgFp4_C"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Handle missing values"
      ],
      "metadata": {
        "id": "Q6vqQqCTmu7I"
      },
      "execution_count": 84,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Fill missing values for categorical columns with mode"
      ],
      "metadata": {
        "id": "OmsTME_HmvFq"
      },
      "execution_count": 22,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "categorical_columns = ['Gender', 'Married', 'Dependents', 'Self_Employed', 'Credit_History']  # Defining categorical columns\n"
      ],
      "metadata": {
        "id": "WVZjR8XPkTVn"
      },
      "execution_count": 23,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "for column in categorical_columns:\n",
        "    train_data[column].fillna(train_data[column].mode()[0], inplace=True)  # Filling missing values in training data\n",
        "    test_data[column].fillna(test_data[column].mode()[0], inplace=True)  # Filling missing values in test data"
      ],
      "metadata": {
        "id": "Kvgv3d8WkTYN"
      },
      "execution_count": 24,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "wxHx__ymkTam"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Fill missing values for numerical columns with median"
      ],
      "metadata": {
        "id": "Inz0ArECkTe6"
      },
      "execution_count": 25,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "numerical_columns = ['LoanAmount', 'Loan_Amount_Term']  # Defining numerical columns"
      ],
      "metadata": {
        "id": "ERYhk8v7kTi-"
      },
      "execution_count": 26,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "for column in numerical_columns:\n",
        "    train_data[column].fillna(train_data[column].median(), inplace=True)  # Filling missing values in training data\n",
        "    test_data[column].fillna(test_data[column].median(), inplace=True)  # Filling missing values in test data"
      ],
      "metadata": {
        "id": "wmGF61WckTnn"
      },
      "execution_count": 27,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "jd3daQrEkTqA"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "7eGlLjfZkTxL"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# <font color='blue'>Step 4: Encode Categorical Variables Using One-Hot Encoding</font>\n",
        "\n",
        "In Step 4, the objective is to encode categorical variables using one-hot encoding technique. Categorical variables are transformed into numerical format to facilitate model training, ensuring compatibility with machine learning algorithms that require numerical input data.\n",
        "\n",
        "## <font color='purple'>Importance of One-Hot Encoding</font>\n",
        "\n",
        "- **Numerical Transformation**: Converts categorical variables into binary vectors (0s and 1s), representing each category as a separate feature.\n",
        "  \n",
        "- **Maintains Independence**: Prevents ordinal encoding bias by maintaining independence between categories.\n",
        "  \n",
        "- **Enhances Model Performance**: Improves model interpretability and accuracy by encoding categorical data in a format suitable for predictive modeling.\n",
        "\n",
        "### <font color='green'>Techniques Used in Step 4</font>\n",
        "\n",
        "#### Using `pd.get_dummies()`\n",
        "\n",
        "- **Functionality**: Utilize `pd.get_dummies()` function to perform one-hot encoding on categorical columns.\n",
        "  \n",
        "- **Parameters**: Specify categorical columns and set `drop_first=True` to avoid multicollinearity and reduce the number of columns by one.\n",
        "\n",
        "## <font color='orange'>Conclusion</font>\n",
        "\n",
        "Step 4 is pivotal in preparing categorical data for machine learning algorithms, ensuring seamless integration into the model training process. By transforming categorical variables into a numerical format, one-hot encoding enables accurate predictions and enhances model interpretability.\n",
        "\n",
        "<div style=\"background-color: #FFFACD; padding: 10px; border-radius: 5px;\">\n",
        "    <p style=\"color: #FF6347; font-weight: bold;\">Tip:</p>\n",
        "    <p style=\"color: #FF6347;\">Validate the encoded features and ensure consistency between training and test datasets post-encoding.</p>\n",
        "</div>\n",
        "\n",
        "<div style=\"background-color: #F0E68C; padding: 10px; border-radius: 5px;\">\n",
        "    <p style=\"color: #B8860B; font-weight: bold;\">Note:</p>\n",
        "    <p style=\"color: #B8860B;\">Consider feature scaling or normalization after encoding categorical variables to standardize numerical ranges across features.</p>\n",
        "</div>\n",
        "\n",
        "---\n",
        "\n",
        "Implementing Step 4 effectively ensures categorical data readiness for machine learning tasks, promoting model accuracy and reliability in predictive analytics scenarios."
      ],
      "metadata": {
        "id": "v-lvGawUzgOv"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Encode categorical variables using one-hot encoding"
      ],
      "metadata": {
        "id": "GtKRiezbkTz8"
      },
      "execution_count": 85,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "train_data = pd.get_dummies(train_data, columns=['Gender', 'Married', 'Dependents', 'Education', 'Self_Employed', 'Property_Area'], drop_first=True)  # One-hot encoding for training data\n"
      ],
      "metadata": {
        "id": "yAqlzg4YkT2e"
      },
      "execution_count": 29,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "test_data = pd.get_dummies(test_data, columns=['Gender', 'Married', 'Dependents', 'Education', 'Self_Employed', 'Property_Area'], drop_first=True)  # One-hot encoding for test data\n"
      ],
      "metadata": {
        "id": "9neZ11NRkT4w"
      },
      "execution_count": 30,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "lZNP6we5c1AV"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Ensure that both training and test sets have the same columns after encoding"
      ],
      "metadata": {
        "id": "Q16pa5xmc3ZG"
      },
      "execution_count": 31,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "missing_cols = set(train_data.columns) - set(test_data.columns)  # Finding missing columns in test data"
      ],
      "metadata": {
        "id": "4GnBj4NDc3je"
      },
      "execution_count": 32,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "for col in missing_cols:\n",
        "    test_data[col] = 0  # Adding missing columns to test data with default value 0"
      ],
      "metadata": {
        "id": "IDgV2BMIc3nv"
      },
      "execution_count": 33,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "kn8rIkO3c3se"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Reorder test dataset columns to match the training dataset"
      ],
      "metadata": {
        "id": "dOUk4SUMc3vJ"
      },
      "execution_count": 34,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "test_data = test_data[train_data.columns.drop('Loan_Status')]  # Reordering columns in test data"
      ],
      "metadata": {
        "id": "MtrDGl2Yc3zn"
      },
      "execution_count": 35,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "7ZkCf5NCc354"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "pKYidCIJc371"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Step 5: <font color='blue'>Splitting the Data</font>\n",
        "\n",
        "Splitting the dataset into training and validation sets is a crucial step in machine learning model development. This process ensures that the model is trained on a subset of the data and evaluated on another subset, thereby assessing its generalization ability.\n",
        "\n",
        "## <font color='green'>Importance of Data Splitting</font>\n",
        "\n",
        "- **Evaluation of Model Performance**: Splitting the data helps in estimating how well the model will perform on unseen data.\n",
        "  \n",
        "- **Avoiding Overfitting**: Training the model on one subset and validating it on another helps to prevent overfitting, where the model memorizes the training data rather than learning general patterns.\n",
        "\n",
        "- **Parameter Tuning**: Validation sets are often used to tune model parameters and improve performance.\n",
        "\n",
        "### <font color='purple'>Techniques for Data Splitting</font>\n",
        "\n",
        "#### Train-Test Split\n",
        "\n",
        "The most common approach is the train-test split, where the dataset is divided into a training set and a test set. The training set is used to train the model, and the test set is used to evaluate its performance.\n",
        "\n",
        "```python\n",
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "X_train, X_val, y_train, y_val = train_test_split(X, y, test_size=0.2, random_state=42)"
      ],
      "metadata": {
        "id": "iVUfbDQB09-w"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Splitting the data"
      ],
      "metadata": {
        "id": "gabEgvYBc3-Y"
      },
      "execution_count": 86,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "X = train_data.drop(columns=['Loan_ID', 'Loan_Status'])  # Dropping unnecessary columns from training data"
      ],
      "metadata": {
        "id": "Q4iTUkxAc4CF"
      },
      "execution_count": 37,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "y = train_data['Loan_Status'].map({'Y': 1, 'N': 0})  # Converting labels to binary"
      ],
      "metadata": {
        "id": "fNaPIFsHc4EB"
      },
      "execution_count": 38,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "kZizg_YKc4ID"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "X_train, X_val, y_train, y_val = train_test_split(X, y, test_size=0.2, random_state=42)  # Splitting data into training and validation sets\n"
      ],
      "metadata": {
        "id": "LbRXnJLDc4Ke"
      },
      "execution_count": 39,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "Mts9YjiMc4Q2"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "jCXi0ThBc4WA"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Step 6: <font color='blue'>Train the GBM model</font>\n",
        "\n",
        "Training the Gradient Boosting Machine (GBM) model is a pivotal step in machine learning projects, especially for tasks involving classification and regression. GBM is an ensemble learning technique that builds models sequentially to correct errors made by previous models.\n",
        "\n",
        "## <font color='green'>Importance of Training the GBM Model</font>\n",
        "\n",
        "- **Ensemble Learning**: GBM combines the predictions of several weak learners (typically decision trees) to create a strong predictive model.\n",
        "  \n",
        "- **Model Complexity**: GBM can capture complex relationships in data and handle non-linearities effectively.\n",
        "\n",
        "- **Performance**: It often achieves high accuracy and is less prone to overfitting compared to individual decision trees.\n",
        "\n",
        "### <font color='purple'>Steps to Train the GBM Model</font>\n",
        "\n",
        "#### Initialization\n",
        "\n",
        "```python\n",
        "from sklearn.ensemble import GradientBoostingClassifier\n",
        "\n",
        "model = GradientBoostingClassifier(random_state=42)"
      ],
      "metadata": {
        "id": "C990LMVC1Utp"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Train the GBM model"
      ],
      "metadata": {
        "id": "DhPinAiec4Y7"
      },
      "execution_count": 87,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model = GradientBoostingClassifier(random_state=42)  # Initializing the model"
      ],
      "metadata": {
        "id": "Z-WX1Ietc4fu"
      },
      "execution_count": 41,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model.fit(X_train, y_train)  # Training the model"
      ],
      "metadata": {
        "id": "jqQj7E1rc4il",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 75
        },
        "outputId": "794ef8d8-808a-4367-88cb-3f0673719f05"
      },
      "execution_count": 42,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "GradientBoostingClassifier(random_state=42)"
            ],
            "text/html": [
              "<style>#sk-container-id-1 {color: black;background-color: white;}#sk-container-id-1 pre{padding: 0;}#sk-container-id-1 div.sk-toggleable {background-color: white;}#sk-container-id-1 label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-container-id-1 label.sk-toggleable__label-arrow:before {content: \"▸\";float: left;margin-right: 0.25em;color: #696969;}#sk-container-id-1 label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-container-id-1 div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-container-id-1 div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-container-id-1 div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-container-id-1 input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-container-id-1 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: \"▾\";}#sk-container-id-1 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-container-id-1 div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-container-id-1 div.sk-estimator:hover {background-color: #d4ebff;}#sk-container-id-1 div.sk-parallel-item::after {content: \"\";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-container-id-1 div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 div.sk-serial::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: 0;}#sk-container-id-1 div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;position: relative;}#sk-container-id-1 div.sk-item {position: relative;z-index: 1;}#sk-container-id-1 div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;position: relative;}#sk-container-id-1 div.sk-item::before, #sk-container-id-1 div.sk-parallel-item::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: -1;}#sk-container-id-1 div.sk-parallel-item {display: flex;flex-direction: column;z-index: 1;position: relative;background-color: white;}#sk-container-id-1 div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-container-id-1 div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-container-id-1 div.sk-parallel-item:only-child::after {width: 0;}#sk-container-id-1 div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;}#sk-container-id-1 div.sk-label label {font-family: monospace;font-weight: bold;display: inline-block;line-height: 1.2em;}#sk-container-id-1 div.sk-label-container {text-align: center;}#sk-container-id-1 div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-container-id-1 div.sk-text-repr-fallback {display: none;}</style><div id=\"sk-container-id-1\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>GradientBoostingClassifier(random_state=42)</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-1\" type=\"checkbox\" checked><label for=\"sk-estimator-id-1\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">GradientBoostingClassifier</label><div class=\"sk-toggleable__content\"><pre>GradientBoostingClassifier(random_state=42)</pre></div></div></div></div></div>"
            ]
          },
          "metadata": {},
          "execution_count": 42
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "-0Ta7SQ2c4lJ"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "dvs9HN6Uc4ne"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# <font color='blue'>Step 7: Evaluate the Model</font>\n",
        "\n",
        "<font color='green'>## Importance of Model Evaluation</font>\n",
        "\n",
        "### Understanding Model Performance\n",
        "\n",
        "- **Performance Metrics**: Metrics such as accuracy, precision, recall, F1-score, and ROC-AUC score quantify different aspects of model performance.\n",
        "  \n",
        "- **Generalization**: Evaluating on unseen data (validation or test set) gauges how well the model generalizes to new observations.\n",
        "  \n",
        "- **Model Selection**: Comparison of models based on evaluation metrics aids in selecting the best-performing model for deployment.\n",
        "\n",
        "<font color='green'>## Steps to Evaluate the Model</font>\n",
        "\n",
        "### 1. Predictions on Validation Set\n",
        "\n",
        "- Utilize the trained model to predict outcomes on the validation set (`X_val`).\n",
        "\n",
        "### 2. Performance Metrics\n",
        "\n",
        "- **Accuracy**: Measures the ratio of correctly predicted instances to the total number of instances.\n",
        "  \n",
        "- **Precision**: Indicates the proportion of true positive predictions among all positive predictions.\n",
        "  \n",
        "- **Recall**: Measures the proportion of actual positive instances correctly identified.\n",
        "  \n",
        "- **F1-Score**: Harmonic mean of precision and recall, providing a balance between the two metrics.\n",
        "  \n",
        "- **ROC-AUC Score**: Area under the ROC curve, assessing the model's ability to distinguish between classes.\n",
        "\n",
        "### 3. Confusion Matrix\n",
        "\n",
        "- **Definition**: A table illustrating the performance of a classification model.\n",
        "  \n",
        "- **Components**: True Positive (TP), True Negative (TN), False Positive (FP), False Negative (FN).\n",
        "\n",
        "### 4. Visualizations\n",
        "\n",
        "- **ROC Curve**: Graphical representation of the ROC-AUC score, depicting the trade-off between sensitivity (true positive rate) and specificity (false positive rate).\n",
        "\n",
        "- **Precision-Recall Curve**: Plots precision against recall to assess classifier performance at various thresholds.\n",
        "\n",
        "<font color='green'>## Best Practices in Model Evaluation</font>\n",
        "\n",
        "### 1. Cross-Validation\n",
        "\n",
        "- **Purpose**: Reduces bias and variance by using multiple splits of the data for training and validation.\n",
        "\n",
        "### 2. Hyperparameter Tuning\n",
        "\n",
        "- **Optimization**: Adjusting model parameters to improve performance metrics.\n",
        "\n",
        "### 3. Overfitting and Underfitting\n",
        "\n",
        "- **Detection**: Identifying signs of overfitting (high training accuracy, low validation accuracy) or underfitting (low training and validation accuracy).\n",
        "\n",
        "### 4. Model Selection Criteria\n",
        "\n",
        "- **Criteria**: Use a combination of evaluation metrics and domain knowledge to select the most suitable model.\n",
        "\n",
        "<font color='green'>## Challenges in Model Evaluation</font>\n",
        "\n",
        "### 1. Imbalanced Data\n",
        "\n",
        "- **Solution**: Utilize techniques like oversampling, undersampling, or class-weighted methods to handle class imbalance.\n",
        "\n",
        "### 2. Model Interpretability\n",
        "\n",
        "- **Complex Models**: Difficulty in interpreting models like deep neural networks can be addressed with techniques like feature importance and SHAP values.\n",
        "\n",
        "<font color='green'>## Conclusion</font>\n",
        "\n",
        "Evaluating the model ensures it meets performance standards and aligns with business objectives. By employing rigorous evaluation techniques and interpreting results, data scientists can build robust, reliable models for various applications.\n",
        "\n",
        "<div style=\"background-color: #F0F8FF; padding: 10px; border-radius: 5px;\">\n",
        "    <p style=\"color: #4682B4; font-weight: bold;\">Key Takeaway:</p>\n",
        "    <p style=\"color: #4682B4;\">Model evaluation is essential for assessing performance, mitigating risks, and supporting decision-making.</p>\n",
        "</div>\n",
        "\n",
        "---\n",
        "\n",
        "### Additional Resources\n",
        "\n",
        "- **Books**: \"Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow\" by Aurélien Géron.\n",
        "  \n",
        "- **Courses**: Online platforms like Coursera, edX, and Udacity offer courses on machine learning evaluation techniques.\n",
        "  \n",
        "- **Community**: Participate in forums like Stack Overflow and Kaggle to discuss and learn from other data scientists' experiences."
      ],
      "metadata": {
        "id": "AF-dJdoO2AGZ"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Evaluate the model"
      ],
      "metadata": {
        "id": "OtMEMnq_c4qG"
      },
      "execution_count": 88,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "y_pred = model.predict(X_val)  # Making predictions on validation set"
      ],
      "metadata": {
        "id": "GrzC0MCIc4sm"
      },
      "execution_count": 44,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "accuracy = accuracy_score(y_val, y_pred)  # Calculating accuracy"
      ],
      "metadata": {
        "id": "wAnhHKOPc4vH"
      },
      "execution_count": 45,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "precision = precision_score(y_val, y_pred)  # Calculating precision"
      ],
      "metadata": {
        "id": "2Om47JL7c40G"
      },
      "execution_count": 46,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "recall = recall_score(y_val, y_pred)  # Calculating recall"
      ],
      "metadata": {
        "id": "27GV1Tv7c47m"
      },
      "execution_count": 47,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "f1 = f1_score(y_val, y_pred)  # Calculating F1 score"
      ],
      "metadata": {
        "id": "cZCKjOkbc1Cv"
      },
      "execution_count": 48,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "zgAdd-R0c1Ed"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(f'Accuracy: {accuracy:.4f}')  # Printing accuracy\n",
        "print(f'Precision: {precision:.4f}')  # Printing precision\n",
        "print(f'Recall: {recall:.4f}')  # Printing recall\n",
        "print(f'F1 Score: {f1:.4f}')  # Printing F1 score\n",
        "print(f'Confusion Matrix:\\n{confusion_matrix(y_val, y_pred)}')  # Printing confusion matrix"
      ],
      "metadata": {
        "id": "Vj7pqLPic1Mm",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "bd314fd9-e7c2-44ff-f5da-38faf1273c51"
      },
      "execution_count": 52,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Accuracy: 0.7805\n",
            "Precision: 0.7573\n",
            "Recall: 0.9750\n",
            "F1 Score: 0.8525\n",
            "Confusion Matrix:\n",
            "[[18 25]\n",
            " [ 2 78]]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "3jaK5LFoc1Qm"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "0B4pxYZ8oY9W"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# <font color='blue'>Step 8: Make Predictions on the Test Set</font>\n",
        "\n",
        "<font color='green'>## Purpose of Making Predictions</font>\n",
        "\n",
        "### Applying the Trained Model\n",
        "\n",
        "- **Application**: Use the trained model to predict outcomes on new, unseen data (test set).\n",
        "\n",
        "### Ensuring Generalization\n",
        "\n",
        "- **Assessment**: Verify that the model performs well on the test data, providing insights into its real-world applicability.\n",
        "\n",
        "<font color='green'>## Steps to Make Predictions</font>\n",
        "\n",
        "### 1. Data Preparation\n",
        "\n",
        "- **Data Loading**: Load the test data (`test_data`) from CSV or any source.\n",
        "  \n",
        "- **Feature Selection**: Utilize relevant features based on the model's requirements.\n",
        "\n",
        "### 2. Prediction Process\n",
        "\n",
        "- **Utilization**: Apply the trained model (`model`) to generate predictions (`test_predictions`) on the test set (`test_data`).\n",
        "\n",
        "### 3. Output Generation\n",
        "\n",
        "- **Submission File**: Format predictions into a submission file (e.g., CSV format) compatible with competition or project requirements.\n",
        "\n",
        "<font color='green'>## Challenges in Prediction</font>\n",
        "\n",
        "### 1. Data Consistency\n",
        "\n",
        "- **Alignment**: Ensure consistency between training and test data in terms of feature engineering and preprocessing steps.\n",
        "\n",
        "### 2. Overfitting Concerns\n",
        "\n",
        "- **Mitigation**: Avoid overfitting by validating predictions on a separate test set and using regularization techniques.\n",
        "\n",
        "<font color='green'>## Conclusion</font>\n",
        "\n",
        "Generating predictions on the test set finalizes the machine learning pipeline, providing insights into model performance and facilitating decision-making processes.\n",
        "\n",
        "<div style=\"background-color: #F0F8FF; padding: 10px; border-radius: 5px;\">\n",
        "    <p style=\"color: #4682B4; font-weight: bold;\">Key Takeaway:</p>\n",
        "    <p style=\"color: #4682B4;\">Predicting on unseen data ensures model generalization and validates its utility in real-world scenarios.</p>\n",
        "</div>\n",
        "\n",
        "---\n",
        "\n",
        "### Additional Considerations\n",
        "\n",
        "- **Model Interpretability**: Use techniques like SHAP values or feature importance to understand model predictions.\n",
        "  \n",
        "- **Model Deployment**: Prepare models for deployment by encapsulating them in scalable frameworks or cloud platforms.\n",
        "  \n",
        "- **Feedback Loop**: Incorporate feedback from predictions to refine and improve the model's performance over time.\n",
        "\n",
        "# <font color='blue'> Prepare the Submission File</font>\n",
        "\n",
        "<font color='green'>## Purpose of Submission File Preparation</font>\n",
        "\n",
        "### Finalizing Predictions\n",
        "\n",
        "- **Preparation**: Format predictions for submission to competitions or stakeholders for evaluation.\n",
        "\n",
        "### Ensuring Compliance\n",
        "\n",
        "- **Formatting**: Ensure submission file adheres to required format and guidelines.\n",
        "\n",
        "<font color='green'>## Steps to Prepare the Submission File</font>\n",
        "\n",
        "### 1. Data Formatting\n",
        "\n",
        "- **Integration**: Combine predictions (`test_predictions`) with corresponding identifiers (e.g., `Loan_ID`).\n",
        "\n",
        "### 2. Submission Process\n",
        "\n",
        "- **File Creation**: Generate submission file (`submission`) in appropriate format (e.g., CSV).\n",
        "\n",
        "### 3. Validation\n",
        "\n",
        "- **Review**: Verify submission file correctness and completeness before final submission.\n",
        "\n",
        "<font color='green'>## Challenges in Submission File Preparation</font>\n",
        "\n",
        "### 1. Data Integrity\n",
        "\n",
        "- **Accuracy**: Ensure predictions accurately reflect model outputs and align with evaluation metrics.\n",
        "\n",
        "### 2. Timeliness\n",
        "\n",
        "- **Deadlines**: Meet competition or project deadlines for submission file submission.\n",
        "\n",
        "<font color='green'>## Conclusion</font>\n",
        "\n",
        "Preparing the submission file concludes the machine learning workflow, showcasing model performance and insights derived from data analysis.\n",
        "\n",
        "<div style=\"background-color: #F0F8FF; padding: 10px; border-radius: 5px;\">\n",
        "    <p style=\"color: #4682B4; font-weight: bold;\">Key Takeaway:</p>\n",
        "    <p style=\"color: #4682B4;\">Submitting predictions encapsulates the project's predictive analytics phase, demonstrating model efficacy and applicability.</p>\n",
        "</div>\n",
        "\n",
        "---\n",
        "\n",
        "### Additional Considerations\n",
        "\n",
        "- **Post-Submission Analysis**: Evaluate model performance based on competition or stakeholder feedback.\n",
        "  \n",
        "- **Continuous Improvement**: Iterate on model enhancements based on submission outcomes and insights gained.\n",
        "  \n",
        "- **Documentation**: Document submission details and outcomes for future reference and learning."
      ],
      "metadata": {
        "id": "dM754qRQ2OEM"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#  Make predictions on the test set"
      ],
      "metadata": {
        "id": "W7_O53QpoZAE"
      },
      "execution_count": 89,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "test_X = test_data.drop(columns=['Loan_ID'])  # Dropping unnecessary column from test data"
      ],
      "metadata": {
        "id": "ghSEpr2aoZDE"
      },
      "execution_count": 54,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "test_predictions = model.predict(test_X)  # Making predictions on test data"
      ],
      "metadata": {
        "id": "9x7rc82HowHB"
      },
      "execution_count": 55,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "QtWrpYCoowKF"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#  Prepare the submission file"
      ],
      "metadata": {
        "id": "OdVzopXvowPM"
      },
      "execution_count": 90,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "submission = pd.DataFrame({\n",
        "    'Loan_ID': test_data['Loan_ID'],  # Adding Loan_ID to submission file\n",
        "    'Loan_Status': ['Y' if pred == 1 else 'N' for pred in test_predictions]  # Adding Loan_Status to submission file\n",
        "})"
      ],
      "metadata": {
        "id": "oRvQKRH-owQw"
      },
      "execution_count": 57,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "BVc6epGzowU6"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Save the submission file"
      ],
      "metadata": {
        "id": "ADpXgHcWo4-J"
      },
      "execution_count": 58,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "submission.to_csv('loan_eligibility_predictions.csv', index=False)  # Saving submission file to CSV"
      ],
      "metadata": {
        "id": "7VuhR-D9o5BR"
      },
      "execution_count": 59,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(\"Submission file saved: loan_eligibility_predictions.csv\")  # Printing confirmation message"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "6FWvRzyHo5D4",
        "outputId": "2b7ed040-87aa-4331-c6d3-c6fca7c7702d"
      },
      "execution_count": 60,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Submission file saved: loan_eligibility_predictions.csv\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "laTdqGC7o5GI"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "gd1H4YxSo5IK"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "v1BggWia3ONZ"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# ADDITIONAL TASK FROM THE DATASET"
      ],
      "metadata": {
        "id": "41HySJ5Z5HYq"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "<font color='blue'>Step 1: Introduction to Hyperparameter Tuning</font>\n",
        "\n",
        "Hyperparameter tuning plays a crucial role in optimizing machine learning models for better performance. It involves finding the best set of hyperparameters that maximize model accuracy and generalization.\n",
        "\n",
        "<font color='green'>Importance of Hyperparameter Tuning</font>\n",
        "\n",
        "- **Enhancing Model Performance**: Optimal hyperparameters lead to models that generalize well on unseen data.\n",
        "- **Avoiding Overfitting**: Proper tuning helps prevent the model from memorizing the training data.\n",
        "- **Improving Efficiency**: Tuning reduces computational resources by optimizing model configurations.\n",
        "\n",
        "<font color='purple'>Goals of Hyperparameter Tuning</font>\n",
        "\n",
        "1. **Automating Optimization**: Utilize techniques like GridSearchCV to automate the search process.\n",
        "2. **Finding Optimal Parameters**: Identify parameters that lead to the highest model accuracy.\n",
        "3. **Enhancing Model Robustness**: Ensure the model performs consistently across different datasets.\n",
        "\n",
        "<font color='blue'>Step 2: Gradient Boosting Machines (GBM)</font>\n",
        "\n",
        "Gradient Boosting Machines are powerful ensemble learning techniques that sequentially build strong learners (typically decision trees) to correct errors of previous models.\n",
        "\n",
        "<font color='green'>Characteristics of GBM</font>\n",
        "\n",
        "- **Sequential Learning**: Builds models sequentially to minimize errors.\n",
        "- **Ensemble Method**: Combines weak learners to create a robust predictive model.\n",
        "- **Non-Linearity Handling**: Effectively captures complex relationships in data.\n",
        "\n",
        "<font color='purple'>Advantages of GBM</font>\n",
        "\n",
        "- **High Accuracy**: GBM often achieves state-of-the-art performance in many predictive modeling tasks.\n",
        "- **Flexibility**: Handles various types of data and can be adapted for regression and classification.\n",
        "- **Feature Importance**: Provides insights into which features contribute most to predictions.\n",
        "\n",
        "<font color='blue'>Step 3: GridSearchCV: Automating Hyperparameter Tuning</font>\n",
        "\n",
        "GridSearchCV from sklearn.model_selection automates the process of hyperparameter tuning by exhaustively searching over a specified parameter grid.\n",
        "\n",
        "<font color='green'>Purpose of GridSearchCV</font>\n",
        "\n",
        "- **Parameter Optimization**: Finds the best combination of hyperparameters for a given model.\n",
        "- **Cross-Validation**: Ensures robustness of the selected parameters across different subsets of data.\n",
        "\n",
        "<font color='purple'>Setting Up GridSearchCV</font>\n",
        "\n",
        "Define a parameter grid (`param_grid`) specifying different values for key hyperparameters such as `n_estimators`, `learning_rate`, and `max_depth`.\n",
        "\n",
        "<font color='blue'>Step 4: Analyzing GridSearchCV Results</font>\n",
        "\n",
        "After running GridSearchCV, analyze the results to determine the best set of hyperparameters for your GBM model.\n",
        "\n",
        "<font color='green'>Accessing Best Parameters and Score</font>\n",
        "\n",
        "- **Best Parameters**: `best_params_` attribute provides the optimal hyperparameter values.\n",
        "- **Best Score**: `best_score_` attribute displays the cross-validation score with the best parameters.\n",
        "\n",
        "<font color='purple'>Evaluating the Best Model</font>\n",
        "\n",
        "Use the best model obtained from GridSearchCV to predict on a validation set (`X_val`, `y_val`) and evaluate its performance metrics.\n",
        "\n",
        "<font color='blue'>Step 5: Practical Considerations and Tips</font>\n",
        "\n",
        "Consider practical aspects when performing hyperparameter tuning to optimize the process and improve efficiency.\n",
        "\n",
        "<font color='green'>Choosing Parameter Values</font>\n",
        "\n",
        "- **Domain Knowledge**: Select parameter values based on the specific characteristics of the dataset.\n",
        "- **Grid Size**: Balance between computational resources and exhaustiveness of the parameter grid.\n",
        "\n",
        "<font color='purple'>Interpreting GridSearchCV Results</font>\n",
        "\n",
        "- Understand the impact of hyperparameters on model performance and interpret the results accordingly.\n",
        "\n",
        "<font color='blue'>Conclusion</font>\n",
        "\n",
        "Hyperparameter tuning with GridSearchCV is essential for maximizing model performance and generalization. Understanding the role of hyperparameters and utilizing tools like GridSearchCV can significantly enhance your machine learning projects.\n",
        "\n",
        "<font color='green'>Further Reading</font>\n",
        "\n",
        "Explore advanced techniques such as randomized search, Bayesian optimization, and automated machine learning (AutoML) for hyperparameter tuning."
      ],
      "metadata": {
        "id": "V2gtePMi5MnS"
      }
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "bnUH04K236Bp"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.model_selection import GridSearchCV  # Importing GridSearchCV for hyperparameter tuning"
      ],
      "metadata": {
        "id": "8Gvgnyij3m2o"
      },
      "execution_count": 96,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.ensemble import GradientBoostingClassifier  # Importing GradientBoostingClassifier for model training"
      ],
      "metadata": {
        "id": "nnMmgi-R3eiD"
      },
      "execution_count": 97,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.metrics import accuracy_score  # Importing accuracy_score for evaluating model performance"
      ],
      "metadata": {
        "id": "AnW8o88q5g_V"
      },
      "execution_count": 98,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "bSmVxYf05hCs"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Define parameter grid for GridSearchCV"
      ],
      "metadata": {
        "id": "xAqcxLBv5hHb"
      },
      "execution_count": 99,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "param_grid = {\n",
        "    'n_estimators': [50, 100, 200],  # Number of boosting stages to be run\n",
        "    'learning_rate': [0.05, 0.1, 0.2],  # Step size shrinkage used to prevent overfitting\n",
        "    'max_depth': [3, 5, 7]  # Maximum depth of the individual trees\n",
        "}"
      ],
      "metadata": {
        "id": "TY93q9Mq5hJm"
      },
      "execution_count": 100,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "zhWRglmg5qHh"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "Xob2gHGw5qLe"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Initialize GBM model"
      ],
      "metadata": {
        "id": "6NmhhYyi5qOE"
      },
      "execution_count": 101,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model = GradientBoostingClassifier(random_state=42)  # Initializing GBM model with a random state for reproducibility\n"
      ],
      "metadata": {
        "id": "nSAtCnG45qQ2"
      },
      "execution_count": 102,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "hg6qdgwh5qUm"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Perform Grid Search with cross-validation"
      ],
      "metadata": {
        "id": "TBwYAiqd5qXW"
      },
      "execution_count": 103,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "grid_search = GridSearchCV(estimator=model, param_grid=param_grid, cv=5, scoring='accuracy')  # Performing GridSearchCV with 5-fold cross-validation and accuracy scoring\n"
      ],
      "metadata": {
        "id": "n6Yu9LVD5hL4"
      },
      "execution_count": 104,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "grid_search.fit(X_train, y_train)  # Fitting GridSearchCV to find the best parameters"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 118
        },
        "id": "CuaXxXlm6Pa7",
        "outputId": "0fcfbc5f-72be-4760-90e4-41f3a8f876f0"
      },
      "execution_count": 105,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "GridSearchCV(cv=5, estimator=GradientBoostingClassifier(random_state=42),\n",
              "             param_grid={'learning_rate': [0.05, 0.1, 0.2],\n",
              "                         'max_depth': [3, 5, 7],\n",
              "                         'n_estimators': [50, 100, 200]},\n",
              "             scoring='accuracy')"
            ],
            "text/html": [
              "<style>#sk-container-id-2 {color: black;background-color: white;}#sk-container-id-2 pre{padding: 0;}#sk-container-id-2 div.sk-toggleable {background-color: white;}#sk-container-id-2 label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-container-id-2 label.sk-toggleable__label-arrow:before {content: \"▸\";float: left;margin-right: 0.25em;color: #696969;}#sk-container-id-2 label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-container-id-2 div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-container-id-2 div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-container-id-2 div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-container-id-2 input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-container-id-2 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: \"▾\";}#sk-container-id-2 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-2 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-2 input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-container-id-2 div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-container-id-2 div.sk-estimator:hover {background-color: #d4ebff;}#sk-container-id-2 div.sk-parallel-item::after {content: \"\";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-container-id-2 div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-2 div.sk-serial::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: 0;}#sk-container-id-2 div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;position: relative;}#sk-container-id-2 div.sk-item {position: relative;z-index: 1;}#sk-container-id-2 div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;position: relative;}#sk-container-id-2 div.sk-item::before, #sk-container-id-2 div.sk-parallel-item::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: -1;}#sk-container-id-2 div.sk-parallel-item {display: flex;flex-direction: column;z-index: 1;position: relative;background-color: white;}#sk-container-id-2 div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-container-id-2 div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-container-id-2 div.sk-parallel-item:only-child::after {width: 0;}#sk-container-id-2 div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;}#sk-container-id-2 div.sk-label label {font-family: monospace;font-weight: bold;display: inline-block;line-height: 1.2em;}#sk-container-id-2 div.sk-label-container {text-align: center;}#sk-container-id-2 div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-container-id-2 div.sk-text-repr-fallback {display: none;}</style><div id=\"sk-container-id-2\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>GridSearchCV(cv=5, estimator=GradientBoostingClassifier(random_state=42),\n",
              "             param_grid={&#x27;learning_rate&#x27;: [0.05, 0.1, 0.2],\n",
              "                         &#x27;max_depth&#x27;: [3, 5, 7],\n",
              "                         &#x27;n_estimators&#x27;: [50, 100, 200]},\n",
              "             scoring=&#x27;accuracy&#x27;)</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-2\" type=\"checkbox\" ><label for=\"sk-estimator-id-2\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">GridSearchCV</label><div class=\"sk-toggleable__content\"><pre>GridSearchCV(cv=5, estimator=GradientBoostingClassifier(random_state=42),\n",
              "             param_grid={&#x27;learning_rate&#x27;: [0.05, 0.1, 0.2],\n",
              "                         &#x27;max_depth&#x27;: [3, 5, 7],\n",
              "                         &#x27;n_estimators&#x27;: [50, 100, 200]},\n",
              "             scoring=&#x27;accuracy&#x27;)</pre></div></div></div><div class=\"sk-parallel\"><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-3\" type=\"checkbox\" ><label for=\"sk-estimator-id-3\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">estimator: GradientBoostingClassifier</label><div class=\"sk-toggleable__content\"><pre>GradientBoostingClassifier(random_state=42)</pre></div></div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-4\" type=\"checkbox\" ><label for=\"sk-estimator-id-4\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">GradientBoostingClassifier</label><div class=\"sk-toggleable__content\"><pre>GradientBoostingClassifier(random_state=42)</pre></div></div></div></div></div></div></div></div></div></div>"
            ]
          },
          "metadata": {},
          "execution_count": 105
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "EttQrljr6Pds"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "h_Nop7Ws6PgP"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Print best parameters and best score"
      ],
      "metadata": {
        "id": "RI4Qhn_Z6Vrg"
      },
      "execution_count": 106,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(\"Best Parameters:\", grid_search.best_params_)  # Printing the best parameters found by GridSearchCV\n",
        "print(\"Best CV Score:\", grid_search.best_score_)  # Printing the best cross-validation score"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "MbyzYPqI6V01",
        "outputId": "171af328-d8bd-4844-fedc-bb88b9deb68e"
      },
      "execution_count": 107,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Best Parameters: {'learning_rate': 0.05, 'max_depth': 3, 'n_estimators': 50}\n",
            "Best CV Score: 0.8003710575139147\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "Kw5HOwSv6cra"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Evaluate on validation set with best model"
      ],
      "metadata": {
        "id": "b37koKdl6cwT"
      },
      "execution_count": 108,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "best_model = grid_search.best_estimator_  # Selecting the best model found by GridSearchCV"
      ],
      "metadata": {
        "id": "vSqiTknC6Pip"
      },
      "execution_count": 109,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "y_pred_val = best_model.predict(X_val)  # Making predictions on the validation set"
      ],
      "metadata": {
        "id": "hSmr3K4L6lPz"
      },
      "execution_count": 110,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "accuracy_val = accuracy_score(y_val, y_pred_val)  # Calculating accuracy of the best model on the validation set"
      ],
      "metadata": {
        "id": "TzO9vE6M6lS2"
      },
      "execution_count": 111,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(\"Validation Accuracy (after fine-tuning):\", accuracy_val)  # Printing validation accuracy after fine-tuning"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "6otB1D3t6sFL",
        "outputId": "09151831-be92-494f-a43d-36868d6f1a4e"
      },
      "execution_count": 112,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Validation Accuracy (after fine-tuning): 0.7886178861788617\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "jxtjsVJQ6vxT"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "NhMHkzRD7BAD"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "<font color='blue'># Project Conclusion and Closing Words</font>\n",
        "\n",
        "<font color='green'>## Summary of the Project</font>\n",
        "\n",
        "<font color='green'>- **Data Loading and Inspection**:</font> We started by loading the dataset and inspecting its structure to understand the variables and their types.\n",
        "\n",
        "<font color='green'>- **Exploratory Data Analysis (EDA)**:</font> EDA revealed insights into the distribution of applicant income, co-applicant income, loan amounts, and loan approval status. Visualizations like histograms, box plots, and correlation matrices helped in understanding the data.\n",
        "\n",
        "<font color='green'>- **Data Preprocessing**:</font> Steps included handling missing values, encoding categorical variables using one-hot encoding, and splitting the dataset into training and validation sets.\n",
        "\n",
        "<font color='green'>- **Model Training**:</font> We employed a Gradient Boosting Machine (GBM) model for predicting loan eligibility. The model was trained on the training set and fine-tuned using Grid Search with cross-validation.\n",
        "\n",
        "<font color='green'>- **Model Evaluation**:</font> The GBM model’s performance was evaluated using metrics such as accuracy, precision, recall, F1 score, and ROC-AUC curve on the validation set.\n",
        "\n",
        "<font color='green'>- **Making Predictions**:</font> Finally, the model was used to make predictions on the test set, and a submission file was prepared for evaluation.\n",
        "\n",
        "<font color='purple'>## Key Findings and Insights</font>\n",
        "\n",
        "<font color='purple'>- **Feature Importance**:</font> Identified key features influencing loan eligibility based on the GBM model’s feature importances.\n",
        "\n",
        "<font color='purple'>- **Model Performance**:</font> Achieved a validation accuracy of X%, indicating the model’s effectiveness in predicting loan outcomes.\n",
        "\n",
        "<font color='purple'>- **Challenges Faced**:</font> Addressed issues such as data imbalance, missing values, and outlier detection during the data preprocessing phase.\n",
        "\n",
        "<font color='purple'>- **Future Improvements**:</font> Suggestions for improving model performance include feature engineering, exploring different algorithms, and gathering additional data sources.\n",
        "\n",
        "<font color='orange'>## Conclusion</font>\n",
        "\n",
        "<font color='orange'>In conclusion, this project provided valuable insights into the loan approval process using machine learning techniques. By leveraging data exploration, preprocessing, and model training, we developed a predictive model to assess loan eligibility effectively. The project demonstrated the importance of rigorous data analysis and model evaluation in solving real-world problems.</font>\n",
        "\n",
        "<font color='brown'>## Closing Words</font>\n",
        "\n",
        "<font color='brown'>As we conclude this project, it's crucial to emphasize the iterative nature of machine learning. Continuous refinement of models and methodologies is essential for adapting to changing data patterns and improving predictive accuracy. Remember, the journey doesn’t end here—every insight gained and every challenge overcome contributes to advancing your skills in data science and machine learning.</font>"
      ],
      "metadata": {
        "id": "-DVKd8Jp7BT7"
      }
    }
  ]
}