This project contains notebook and final report that explains all the steps followed for evaluating performance of RIPPER and Hoeffding Tree Algorithm in solving patient no-show problem. For this project, SHAP (SHapley Additive exPlanations) has been used for the feature selection despite the fact that this method does not have any correlation with the methods used for prediction but, I wanted to study, how this method helps us in achieving interpretability in feature selection process regardless of the type of predictive models.
I have used Kaggle dataset for this study.