Forecasting
for Disease Biosurveillance.
I build county-level machine learning systems for Highly Pathogenic Avian Influenza (HPAI) detection and one-month-ahead outbreak forecasting, translating environmental and spatial signals into proactive biosurveillance tools.
Forecasting Framework
County-month modeling across the conterminous United States (CONUS), integrating environmental, climatological, and geospatial features with outbreak observations.
Lagged and time-gated feature construction supports one-month-ahead forecasting without forward look-ahead leakage.
Classification and forecasting models are tuned for extreme class imbalance, with recall and balanced accuracy prioritized for disease surveillance.
Feature importance, PCA, calibration checks, and risk tiering make model behavior legible enough for operational decision support.