Drug Classification using Machine Learning and Interpretability

In this paper we have performed drug classification using machine learning models like random forest, decision tree, and logistic regression. To further understand how these models came to their conclusion, we find the interpretability of these models using LIME and SHAP. Towards the end, we found that LIME and SHAP can be used to gain insights into a Machine Learning model and conclude which feature is responsible for the deviation of the results. According to the LIME and SHAP results, Random Forest and Decision Tree ML models are the most suited models for drug classification as both achieved an accuracy of 0.975. Also, Na to K and BP were found to be the most essential features in predicting the outcome.