HCBiLSTM: A hybrid model for predicting heart disease using CNN and BiLSTM algorithms

Abstract

In human existence, healthcare is an inevitable obligation. Heart disease is a broad term that encompasses a variety of illnesses that affect the heart & veins. The early methods for assessing cardiovascular diseases aided in deciding on the progressions that should have occurred in high-risk individuals, therefore lowering their risks. The primary objective is to save human lives by detecting irregularities in heart conditions, which will be accomplished by the identification & processing of raw data derived from cardiac information. The purpose of this article is to construct a hybrid model based on deep learning methods for predicting whether or not a person has heart disease and providing awareness or diagnosis based on that prediction using CNN & Bi-LSTM. We solve the issue of missing data as well as the issue of imbalanced data in the publicly accessible Heart disease Cleveland UCI dataset using data processing methods. We have used an extra tree classifier for feature selection, and CNN-BiLSTM for classification. Experiments have been done on the Heart disease Cleveland UCI dataset which is collected from the Kaggle. The performance of the diagnosis model is obtained by using approaches like classification, accuracy, precision, recall & f1-score. Here a study is done based on the proposed technique and their performance analysis to propose an accurate model of predicting heart disease. Numerous existing methods were examined, with the hybrid model achieving an accuracy level of 96.66%.

Keywords

• Healthcare
• Disease prediction
• Heart disease prediction
• Deep learning
• Classification
• CNN