Hybrid Deep Learning Model for Heart Disease Prediction Using Recurrent Neural Network (RNN)

Abstract

In this paper, we use a recurrent neural network (RNN) that combines multiple gated recurrent units (GRUs), long short-term memory (LSTM), and the Adam optimizer to develop a new hybrid deep learning model for heart disease prediction. This proposed model yielded an excellent accuracy of 98.6876%. This proposed model is a hybrid of GRUs and RNNs model. The model was developed in Python 3.7 by integrating multiple GRUs and RNNs working with Keras and Tensorflow as backends for the deep learning process and is supported by various Python libraries. A recent existing model using RNN achieved 98.23% accuracy, and Deep Neural Network (DNN) achieved 98.5% accuracy. Common drawbacks of existing models are poor accuracy due to the complex design of neural networks, large numbers of neurons with redundancy in neural network models, and Cleveland imbalanced datasets. Experiments were performed using different fitting models, and the results showed that the proposed model using RNN and several GRUs with synthetic minority oversampling technique (SMOTe) achieved the highest level of performance. it was done. This is the most accurate result of his RNN using the Cleveland dataset and shows promise for early prediction of heart disease in patients.