Medical Events Prediction with Missing Features and Imbalanced Classification



Patients in hospitals are often continuously monitored by the medical personnel, who collect data about the patients’ demographics, vital signs and lab test results. In this project, forecasts are made on the future occurrence of medical events such as sepsis, future orders of medical tests, as well as the evolution of key vital signs of patients in the remainder of their stay, based on data available from their first 12 recorded hours of stay. Accurate predictions could potentially help in resource planning and workflow management in hospitals. To address the challenges of missing features and imbalance classification, simple imputer, standard scaler, and histogram-based gradient boosting classification tree are used.

Chenhao Li
Chenhao Li
Reinforcement Learning for Robotics

My research interests focus on the general field of robot learning, including reinforcement learning, developmental robotics and legged intelligence.