Project Showcase
Highlight of my featured works
Nutritional Status Prediction
Predicting nutritional status using demographic and lifestyle factors. Compared Random Forest baseline with Neural Network, focusing on recall for health-related predictions.
🎯 What I Learned
- Applied RobustScaler and StandardScaler to different features
- Trained and compared Random Forest vs Neural Network
- Focused on recall to minimize false negatives in obesity detection
🛠 Tech Stack
- Python, NumPy, Pandas
- Matplotlib, Seaborn, Plotly
- Statsmodels
- Scikit-learn
- TensorFlow / Keras
📌 Evaluation / Next Step
Random Forest achieved higher accuracy (~96%) but had lower recall, which is risky in healthcare settings. Neural Network gave slightly lower accuracy (~88–89%) but better recall, which is more valuable for reducing misclassification of obese individuals. Next steps include oversampling, hyperparameter tuning, and testing ensemble models for improved balance.