Machine Learning Projects
Click an image to open the full PDF report.
Supervised, Unsupervised, Ensemble, AutoML – end‑to‑end solutions
Kaggle Winner: Smoker Prediction
Top‑ranking binary classifier using LightGBM & CatBoost with extensive feature engineering, balanced sampling and SHAP interpretability.
Breast Cancer Classification
Malignant‑vs‑benign detection; multiple models benchmarked, best performer explained with ROC, confusion matrix & feature importance.
Calories Burned Prediction – AutoML
Physical activity dataset fed into AutoML search; optimal regression pipeline selected & tuned autonomously.
Customer Segmentation with RFM & Clustering
K‑Means + hierarchical clustering on RFM metrics to surface high‑value cohorts for targeted marketing.
Body Measurement Prediction
ANSUR II anthropometrics modelled via Random Forest / XGBoost for ergonomic design use‑cases; MAE & RMSE evaluated.
Medical Cost Prediction
Regression analysis of demographic & lifestyle factors on healthcare expenditure; key drivers identified.
Calories Burned EDA & Modeling
Comprehensive EDA followed by Linear, RF, XGBoost regression contest; hyperparameter‑tuned champion highlighted.
Fertilizer Recommendation System
Soil & crop data leveraged to predict optimal N‑P‑K mix; regression + multiclass models underpin agronomic decisions.
Introvert & Extrovert
The Introvert & Extrovert project is a machine learning study aimed at automatically classifying individuals' personalities along the introversion–extroversion spectrum based on behavioral and data-driven features. The focus is on effective feature engineering and leveraging powerful models (e.g., XGBoost) to achieve high prediction accuracy.