Explainable AI for Early Disease Risk Prediction

This paper presented a complete, reproducible XAI pipeline for heart disease risk prediction, from raw data through cleaning, EDA, multi-model training, and SHAP-based interpretability. Four classifiers were evaluated on 1020 cleaned patient records: Random Forest achieved the best standalone accuracy of 0.9510 with AUC 0.9812 and recall 0.9583, verified by 5-fold CV (mean 0.9242 ± 0.0075). SHAP decomposition revealed that ST slope, chest pain type, and exercise angina dominate global predictions — in full agreement with established cardiology. Per-patient waterfall plots make individual predictions readable and challengeable by clinicians. The central finding of this work is that strong classification accuracy and genuine clinical transparency are not competing goals: with SHAP-augmented ensemble learning, both can be achieved simultaneously — and that combination is what will ultimately allow AI-powered risk screening to move from research publication into clinical practice.

[1]   G. A. Roth et al., "Global burden of cardiovascular diseases and risk factors, 1990–2019: Update from the GBD 2019 study," J. Am. Coll. Cardiol., vol. 76, no. 25, pp. 2982–3021, Dec. 2020. doi: 10.1016/j.jacc.2020.11.010 [2]   S. Mohan, C. Thirumalai, and G. Srivastava, "Effective heart disease prediction using hybrid machine learning techniques," IEEE Access, vol. 7, pp. 81542–81554, 2019. doi: 10.1109/ACCESS.2019.2923707 [3]   K. G. Nair and M. Soni, "Heart disease prediction using machine learning algorithms: A review," Int. J. Comput. Appl., vol. 180, no. 40, pp. 1–7, 2018. [4]   A. Holzinger, G. Langs, H. Denk, K. Zatloukal, and H. Müller, "Causability and explainability of artificial intelligence in medicine," WIREs Data Mining Knowl. Discov., vol. 9, no. 4, p. e1312, Jul. 2019. doi: 10.1002/widm.1312 [5]   S. M. Lundberg and S.-I. Lee, "A unified approach to interpreting model predictions," in Proc. 31st Conf. Neural Inf. Process. Syst. (NeurIPS), Long Beach, CA, USA, Dec. 2017, pp. 4765–4774. [6]   A. Javaid et al., "Explainability of machine learning-based cardiovascular risk prediction models: A systematic review," Front. Cardiovasc. Med., vol. 9, p. 884033, May 2022. doi: 10.3389/fcvm.2022.884033 [7]   T. Chen and C. Guestrin, "XGBoost: A scalable tree boosting system," in Proc. 22nd ACM SIGKDD Int. Conf. Knowledge Discovery and Data Mining, San Francisco, CA, USA, Aug. 2016, pp. 785–794. doi: 10.1145/2939672.2939785 [8]   L. Breiman, "Random forests," Mach. Learn., vol. 45, no. 1, pp. 5–32, Oct. 2001. doi: 10.1023/A:1010933404324 [9]   R. Detrano et al., "International application of a new probability algorithm for the diagnosis of coronary artery disease," Am. J. Cardiol., vol. 64, no. 5, pp. 304–310, Aug. 1989. doi: 10.1016/0002-9149(89)90524-9 [10]U. R. Acharya et al., "Application of deep convolutional neural network for automated detection of myocardial infarction using ECG signals," Inf. Sci., vol. 415–416, pp. 190–198, Nov. 2017. doi: 10.1016/j.ins.2017.06.027