Optimized Feature Selection and Classification Framework for Cardiovascular Disease Using Statistical Normalization and Bio-Inspired Algorithms

Heart disease, which includes various circulatory illnesses that largely impact the structure and function of the heart, is a prevalent primary cause of morbidity and mortality. This long-term illness significantly lowers people’s quality of life and raises healthcare costs. This illness will cause the body to react severely, including heart attacks and cardiac arrest, which may outcome in excruciating losses or even death. Early diagnosis is crucial for survival and treatment effectiveness in order to prevent severe consequences. Some medical diagnostic techniques, such as ECG, ECO, and stress tests, are available for the identification of cardiovascular illness; however, these techniques have certain drawbacks and restrictions caused by challenges with cost, personnel, and accessibility. Remedies are thus required to get over these drawbacks, which are satisfied by computer algorithms. Machine learning algorithms meet this requirement by combining various data types, handling metadata, and identifying medical risk factors and patterns that traditional methods are unable to identify. The Ant-Hybrid Support Vector Machine Learning approach, which employs the hybrid kernel support vector machine learning algorithm for disease classification and Ant colony optimization for feature extraction, is used in this research work to enhance the capabilities of feature selection and classification. This technique will improve feature extraction and classification capabilities by identifying intricate patterns in the dataset that are overlooked by current approaches. The performance of the algorithms for the prediction of cardiovascular disease is compared using the different performance metrics such as Accuracy, Precision, Recall and this comparison clearly shows that the proposed model outperforms the existing algorithms with the highest accuracy of 98%. To increase model accuracy, the z-Score Normalization technique is employed during the preprocessing step to normalize the dataset’s values.