Predicting the Conversion from CIS to MS: An Explainable Machine Learning Approach
This study utilizes a public dataset of 273 Mexican mestizo CIS patients with a 10-year follow-up to predict the conversion of Clinically Isolated Syndrome (CIS) to Clinically Definite Multiple Sclerosis (CDMS). The dataset was divided into a training set for cross-validation and feature selection, and a holdout test set for final testing.
An explainable machine learning model, Extreme Gradient Boosting (XGBoost), was employed for the analysis. Feature importance was determined using the SHapley Additive Explanations library (SHAP).
The dataset used in this study can be found here.
A web-based demo of the model is available for testing here.