Diagnosing cardiovascular disease via intelligence in healthcare multimedia: a novel approach

Abstract

Diagnosing cardiovascular disease (CVD) in its early stages remains a challenge despite the existence of all medical technologies and devices that are being used. Besides the digitised form of collecting and organising data, prediction and diagnosis are two stumbling blocks in CVD. This study explores statistical machine learning models with a multimedia health care approach using AI to predict risk factors of heart diseases associated with type 2 diabetes mellitus (T2DM). This study investigates an efficacy of a mathematical model to perform attribute evaluation using information criteria-based selection in LASSO regression. The present study implements the deep learning algorithm using a multilayer perceptron (MLP) classifier with Gaussian process classification (GPC) that provides probabilistic predictions in terms of linear and non-linear functions. The performance of the classifier is evaluated using precision, recall and accuracy metrics. The proposed classification model yields 93.59% accuracy of 10 cross-validations assorted with sigmoid function for better analysis.