Overview
Predicting small molecule effectiveness as Sirtuin 6 (SIRT6) inhibitor candidates based on molecular descriptors and chemical properties.
Metrics & Insights
- Evaluated Decision Trees vs Scaled K-Nearest Neighbors on 100 compounds with 6 chemical descriptors.
- Scaled KNN achieved 85.0% test accuracy and 81.25% cross-validation score, mitigating severe overfitting observed in Decision Trees (100% train vs 80% test).
Tech Stack: Python, Scikit-learn, Streamlit, Pandas, Matplotlib
Last updated on September 16, 2026 at 12:23 PM UTC+7. See Changelog