Optimized ensemble model for accurate prediction of cardiac vascular calcification in diabetic patients.

Abstract

Aim: Cardiovascular diseases (CVD) are a major threat to diabetic patients, with cardiac vascular calcification (CVC) as a key predictive factor. This study seeks to improve the prediction of these calcifications using advanced machine learning (ML) algorithms. However, current ML and Artificial Intelligence (AI) methods face challenges such as limited sample sizes, insufficient data, high time complexity, long processing times, and significant implementation costs.

Method: To predict CVC in diabetic patients, the Simple linear iterative clustering based Ensemble Artificial Neural Network (SLIC-EANN) model is proposed in this paper. In this research article, certain biochemical, imaging, and clinical data are used that are captured from Coronary computed tomography angiography (CCTA) dataset. The proposed model employs preprocessing techniques such as image normalization, image resizing, and image augmentation to clean and simplify the input images. Then Localization of the cardiac vascular calcification is done using the simple linear iterative clustering (SLIC) algorithm. The ensemble artificial neural network (EANN) classifies calcification severity by integrating outputs from three machine learning techniques Support Vector Machine (SVM), Gradient Boosting (GB), and Decision Tree (DT).

Results: This method achieves an accuracy of 98.7% and an error rate of 1.3%, outperforming existing techniques.

Conclusion: A comprehensive analysis is conducted in this research article that concludes that the proposed model achieved better prediction performances of calcification in diabetic patients.