CRISP-DM
- EDA
- Data Cleaning (Detect Missing Values and Outliers, Data Pipeline, Transformation)
- Supervised Machine Learning Modeling (Naive Bayes, Logistic Regression, Neural Network, Random Forest, Gradient Boosting Decision Tree, K-Nearest Neighbors)
- Evaluation (10-Fold Cross-validation, Confusion Matrix [Accuracy, Precision, Recall, Specificity, F1 Score], AUC-ROC Curve, Cumulative Gain and Lift Charts)
- Model Explainability (Relative Feature Importance, Partial Dependencies, SHAP Values)