A heart disease diagnosis method based on structure-adaptive Belief Rule Base and multi-strategy-improved whale optimization algorithm

Heart disease ranks among the leading causes of death worldwide, where early diagnosis and prevention play a critical role in mitigating its impact. The Belief Rule Base (BRB) model, as a semi-quantitative prediction method grounded in expert knowledge, has demonstrated significant potential in diagnosing heart-related conditions. However, determining the BRB model structure based solely on expert knowledge presents substantial challenges in practical diagnostic applications. Therefore, this study introduces a structure-adaptive Belief Rule Base (SA-BRB) for heart disease diagnosis. First, to resolve the reference value selection issue in the BRB model, the K-Means clustering algorithm with an improved initial centroid selection strategy is employed to automatically generate stable reference values. Next, multiple BRB models are constructed using the reference value sets derived from clustering results. Then, the evidential reasoning (ER) approach is utilized for model inference, while a multi-strategy improved whale optimization algorithm (MSIWOA) is developed to optimize model parameters. Finally, a comprehensive evaluation framework is designed to assess model performance across both complexity and accuracy dimensions, enabling flexible adjustments aligned with decision-makers’ requirements. Results from heart disease case studies confirm the effectiveness and practicality of the proposed model.