Automated Cardiovascular Disease Diagnosis Using Honey Badger Optimization With Modified Deep Learning Model

Abstract

The leading cause of death among people around the world is cardiovascular disease (CVD). In order to prevent patients from other damages, precise diagnostics of CVD on time is a crucial factor. Researcher workers are inspired to apply machine learning (ML) for accurate and quick diagnosis of CVD. ML algorithm extracts patterns and hidden relationships in the medical dataset for detecting or predicting disease development. But the prediction of CVD is a challenging task. The increasing size of healthcare dataset has made it a complex task for practitioners to make disease predictions and understand the feature relations. And so, selection of crucial features in a dataset plays a key role in optimizing the performance of ML algorithm. This study develops an Automated Cardiovascular Disease Diagnosis using Honey Badger Optimization with Modified Deep Learning (ACVD-HBOMDL) Model. The major aim of the ACVD-HBOMDL technique lies in the classification of CVD using feature selection (FS) and hyperparameter tuning strategies. Initially, the ACVD-HBOMDL technique applies min-max scaler to preprocess the medical data. To elect an optimal subset of features, the HBO algorithm is used in this work. For CVD classification, deep learning modified neural network (DLMNN) classifier is used and its hyperparameters can be optimally chosen by Bayesian optimization. The experimental results of the ACVD-HBOMDL technique can be tested on benchmark medical dataset and the obtained results demonstrate the significant outcomes of the ACVD-HBOMDL technique over other existing techniques.