Hospitalization-rate-forecasting

The Coronavirus Disease 2019 (COVID-19) pandemic has presented significant challenges for healthcare systems worldwide, necessitating effective strategies for monitoring and predicting hospitalization rates. Accurate short-term forecasting of hospitalization rates can help in resource allocation, emergency preparedness, and informed public health responses (1-2). This study focuses on developing predictive models for weekly COVID-19 hospitalization rates in the state of Maine using machine learning algorithms. In this project, we aim to use a Decision Tree (DT) regression model as a baseline for short-term forecasting of COVID-19 weekly hospitalizations. The simplicity of Decision Trees allows for straightforward interpretability, making them a useful starting point in predictive modeling. To enhance predictive performance and address potential limitations of a single DT model, we will also compare the baseline results with more complex ensemble methods, including Random Forest and XGBoost. These models are well-suited for handling non-linearity and feature interactions, which are common in time series data. The results from this comparative analysis will provide insight into the strengths and weaknesses of using Decision Tree-based ensembles for forecasting hospitalization rates and contribute to a better understanding of model selection for forecasting disease with seasonal patterns.