🍳 What's Cooking? - Kaggle Recipe Cuisine Classification

A machine learning solution for the Kaggle "What's Cooking?" competition that predicts the cuisine of a recipe based on its ingredients.

📝 Description

This repository contains a Jupyter notebook solution for the Kaggle "What's Cooking?" competition. The challenge involves predicting the type of cuisine (e.g., Italian, Chinese, Mexican) based on the list of ingredients in a recipe. The solution uses natural language processing techniques and machine learning algorithms to classify recipes into their respective cuisines.

✨ Features

• 📊 Exploratory Data Analysis - Visualizes cuisine distribution and ingredient frequencies
• 🔍 Ingredient Analysis - Identifies the most common ingredients for different cuisines
• 📋 Text Processing - Transforms ingredient lists into a format suitable for machine learning
• 🧮 Feature Engineering - Uses TF-IDF vectorization to convert text data into numerical features
• 🤖 Machine Learning Model - Implements a Random Forest Classifier for cuisine prediction
• 📈 Prediction & Submission - Generates predictions and creates a submission file for Kaggle

🔧 Prerequisites

To run this notebook, you'll need:

• Python 3.x
• Jupyter Notebook or JupyterLab
• Required Python libraries:
  • pandas
  • numpy
  • matplotlib
  • seaborn
  • scikit-learn

🚀 Usage

1. Clone this repository:

   git clone https://github.com/corticalstack/KaggleWhatsCooking.git
   cd KaggleWhatsCooking

2. Download the competition data from Kaggle and place it in an input directory:

   • train.json - Training data with labeled cuisines
   • test.json - Test data for making predictions

3. Open and run the Jupyter notebook:

   jupyter notebook kernel.ipynb

4. The notebook will:

   • Load and analyze the data
   • Preprocess the ingredients
   • Train a Random Forest model
   • Generate predictions
   • Create a submission.csv file ready for Kaggle submission

📊 Model Performance

The Random Forest Classifier is used with TF-IDF vectorization of ingredients. The model evaluates performance using out-of-bag error estimation.

📜 License

This project is licensed under the MIT License - see the LICENSE file for details.