Predicting Walking Independence Using Machine Learning and Feature Importance Analysis

Key Points

• This research aims to develop a model predicting walking independence at discharge using LightGBM and analyze feature contributions using SHAP.
• Developed a predictive model using Light Gradient Boosting Machine (LightGBM).
• Conducted feature importance analysis through SHAP to assess contributions of various factors.
• Focused on the outcome of walking independence at discharge.
• Identified key features significantly predicting walking independence.
• Demonstrated the effectiveness of LightGBM for this predictive task.
• Showed that certain variables contributed more than others to the prediction model.