COVID-19 Hospital Demand Forecasting
Advanced time-series forecasting models to predict ICU and ventilator demand during peak pandemic waves.
The sudden surges in hospital admissions during the COVID-19 pandemic severely strained healthcare infrastructure. This project focused on developing robust, short-term forecasting models to predict resource requirements, specifically intensive care unit (ICU) beds and mechanical ventilators.
Methodology
Using historical admission rates, regional infection metrics, and demographic data, we applied a combination of ARIMA models and advanced ensemble machine learning techniques (like XGBoost) to generate rolling 7-day and 14-day forecasts.
Impact
These models provided hospital administrators with critical lead time to reallocate resources and staff, significantly optimizing patient care logistics during critical shortage periods.