Work

Evaluating the trade-off between model accuracy, explainability, and fairness in cardiovascular risk prediction

Year: 2026
Type: article
Abstract: Cardiovascular diseases have always been one of the major contributors to deaths worldwide. The modern era ofhealthcare demands the prediction of such risks, which is the need of the hour in the field of medicine and healthcare.Machine learning has proven to be highly accurate in the field of cardiovascular risk prediction. However, the applicationof machine learning in such scenarios also poses many concerns. The highly accurate machine learning model mightturn out to be a black box, and the inability of the machine learning model to provide explanations for its predictionsmight lead to a lack of trust in the model. Fairness in machine learning also affects the performance of the machinelearning model in different groups. The objective of the paper is to analyse the trade-off between the accuracy, fairness,and explainability of machine learning in cardiovascular risk prediction scenarios. Various machine learning modelshave been compared in the context of the cardiovascular domain. The performance of the machine learning model hasbeen analysed in terms of fairness and explainability using feature attribution methods. The results of the experimentshow that although the complex machine learning model is highly accurate in the context of the cardiovascular domain,it is not highly interpretable and is not equally fair in the context of different demographic groups.
Source: PRATIBODH
Language: English
Open Access status: diamond
OpenAlex ID: W7215244869
FWCI:
Cites: 9
Cited by:
Related to: 10