Decision Tree-Based Explainable AI for Diagnosis of Chronic Kidney Disease

• Published in: 2023 5th International Conference on Inventive Research in Computing Applications (ICIRCA) pp. 947 - 952
• Main Authors: N, Manju V, N, Aparna, K, Krishna Sowjanya
• Format: Conference Proceeding
• Language: English
• Published: IEEE 03.08.2023
• Subjects: Chronic kidney disease; Chronic Kidney Disease (CKD); Cross-Validation; Data models; Decision making; Decision Tree based Explainable AI(DT-EAI); Decision trees; Patient monitoring; Real-time systems; Support vector machines
• DOI: 10.1109/ICIRCA57980.2023.10220774

Chronic Kidney Disease (CKD) is a high-risk health condition that is progressive and life-threatening. Early diagnosis of the same is highly recommended and there have been various means, but Explainable AI in CKD is to enhance transparency, trust, and clinical decision-making, The proposed system which is based on Decision Tree based Explainable AI(DT-EAI), aims at providing a confident solution for early diagnosis. It follows a Data-driven approach where the system uses the CKD Data set and preprocess and selects the features using the Gini Importance value, generates a model using the Decision Tree, Interprets the model using the SHAP value, Evaluates and validates using the Cross-Validation and this is iteratively carried out while getting the feedback form health professions on the results and interpretations. The model is refined and enhanced for accuracy and interpretability. The system is later deployed and used for early diagnosis. The performance of the proposed system is evaluated using the F1 score and Fidelity Accuracy Index(FAI).