A collection of my work in data science and machine learning
Explored IMS vs. DMS trade-offs in probabilistic long-term and short-term time series forecasting. Combined probabilistic and point LTSF methods, demonstrating strong performance with quantile-based techniques.
Designed a data-driven simulation for targeted vaccination campaigns with interactive dashboards.
Integrated B-Cos networks into XceptionNet for deepfake detection and proposed the Mask Pointing Game for evaluating explanation quality in interpretable AI.
Modeled flight prices using XGBoost, RandomForest, and SVMs.
Integrated football data from multiple sources using web scraping and schema matching in Java.
Applied DeepAR for probabilistic influenza forecasting, outperforming traditional and neural baselines in accuracy and uncertainty quantification.
Scraped football player and team data.
Compared LASSO and Ridge for feature selection in regression tasks.