Holiday Package Purchase Prediction

Tools: Python, Scikit-learn, Random Forest, Pandas, NumPy, Matplotlib, Seaborn

• Built an end-to-end binary classification model using Random Forest to predict customer holiday package purchase behaviour — data cleaning, feature encoding, EDA, model training, and evaluation.
• Evaluated model using accuracy, precision, recall, F1-score and ROC-AUC; applied feature importance analysis to identify top predictors of customer purchase behaviour.

View code on GitHub

Feature importance

Top predictors identified via Random Forest feature importance.

ROC Curve

Model ROC curve with AUC to demonstrate discriminative ability.