This project aims to develop a model to accurately predict patient length of stay (LoS) in hospitals, enabling more effective resource planning, improved patient care, and enhanced operational efficiency while controlling hospital costs. Current LoS prediction approaches suffer from several limitations: they are often ad-hoc, overly specific to individual hospital environments, and lack standardization in data preprocessing and model tuning. This restricts their generalizability and prevents meaningful comparison between different prediction methods.
The goal would be to develop a robust model that uses routinely collected hospital data of real patients that can be implemented across different hospital settings.