Spline Regressive Quadratic Emphasis Boosted Ensemble Classifier For Heart Disease Prediction

Abstract

Heart disease is a major global health concern affects the heart and blood arteries like arrhythmias, heart failure, and coronary artery disease. Heart diseases are the leading cause of the severe mortality rate worldwide for both men and women. Early detection of heart disease plays a vital role for timely treatment and continuous monitoring by healthcare providers, and reducing mortality rates. Numerous conventional machine learning methods are developed to identify the heart disease over the decades. However, these models faced the challenges of accurate prediction with minimal time consumption. In order to enhance the heart disease prediction accuracy, a novel method called Spline Regressive Quadratic Emphasis Boosting Classifier (SRQE-Boost) model is proposed. The main aim of proposed SRQE-Boost model is to perform the heart disease prediction through the significant feature selection and classification to minimize the time as well as the space consumption. The proposed SRQE-Boost model comprises four processes, namely data acquisition, preprocessing, feature selection and classification. The data acquisition process is the first step for predicting heart disease with large volume of patient data collected from the input dataset. After data acquisition, preprocessing is carried out to minimize the time as well as memory consumption. During data preprocessing, missing data handling using linear spline interpolation method and outlier removal based on Peirce criterion are carried out to organize the dataset into a suitable format. Followed by, feature selection process is employed using factor regressive analysis to select the relevant features to improve the heart disease prediction by minimizing the dimensionality of the dataset. Factor regressive analysis is a type of statistical analysis used for data analysis through measuring the relationships between features and the target based on Tanimoto Similarity Index. Finally, Quadratic Discriminant Emphasis Boosting ensemble classifier is employed for predicting the heart disease presence or absence with the selected features. In this way, accurate heart disease prediction results are observed with minimal time consumption. Experimental evaluation is carried out on performance metrics like accuracy, precision, recall, F1 score, specificity, AUC, MCC, Prediction time, memory consumption, with respect to number of data samples and features. Quantitative analysis results indicate that the proposed SRQE-Boost model achieved better accuracy in disease prediction, and minimizes time as well as memory consumption compared to existing methods