P-HrDPS: Security-Aware Heart Disorder Prediction Support Model Using Ensemble Learning Technique

Biomedical applications' rapid technological development produces vast amounts of personal data. This biological data creates privacy concerns as it discloses private information like a person's health status and way of life. In Biomedical field, heart disorder is the most critical and major threat for the population across the world. Due to the advancement in machine learning and parametric technologies, heart disorder is possible to predict. In this presented work, heart disorder prediction is performed based on univariate and multivariate feature analysis. Along with this, the proposed system keeps private data safe by using Paillier homomorphic encryption technique to encrypt and decrypt the patients' records. In this work, the max voting ensemble learning technique is used by incorporating five machine learning techniques such as support vector machine, logistic regression, naïve Bayes, decision tree and K-nearest neighbor. The outcome shows that ensemble learning performed better in comparison to others. According to the univariate feature analysis, type (0) chest pain is critical than others. In multivariate analysis, chest pain, exang, and old peak parameters are more correlated with the heart disorder than others.