A Novel Early Detection and Prevention of Coronary Heart Disease Framework Using Hybrid Deep Learning Model and Neural Fuzzy Inference System

• Published in: IEEE access 2024, Vol.12, p.26683-26695
• Main Authors: Ramesh, B., Lakshmanna, Kuruva
• Format: Article
• Language: English
• Subjects: Convolutional neural networks; Coronary arteriosclerosis; coronary heart disease; Deep learning; detection; Detection algorithms; Diabetes; Diseases; fuzzy inference systems; Fuzzy logic; Fuzzy systems; Heart; Heart diseases; Inference; Inference algorithms; Long short term memory; Machine learning; optimal scrutiny boosted graph convolutional LSTM; Prediction algorithms; Predictive models; prevention; Training
• ISSN: 2169-3536; 2169-3536
• DOI: 10.1109/ACCESS.2024.3366537

Diabetes is the "mother of all diseases" as it affects multiple organs of body of an individual in some way. Its timely detection and management are critically important. Otherwise, the long run, it can cause several complications in a diabetic. Heart disease is one of the major complications of diabetes.This work proposed an Optimal Scrutiny Boosted Graph Convolutional LSTM (O-SBGC-LSTM), SBGC-LSTM enhanced by Eurygaster Optimization Algorithm (EOA) to tune hyperparameters for early prevention and detection of diabetes disease. This work proposed an Optimal Scrutiny Boosted Graph Convolutional LSTM (O-SBGC-LSTM), SBGC-LSTM enhanced by Eurygaster Optimization Algorithm (EOA) to tune hyperparameters for early prevention and detection of diabetes disease. This method not only captures discriminative features in spatial configuration and temporal dynamics but also explore the co-occurrence relationship between spatial and temporal domains. This method also presents a temporal hierarchical architecture to increase temporal receptive fields of top SBGC-LSTM layer, which boosts the ability to learn high-level semantic representation and significantly reduces computation cost. The performance of O-SBGC-LSTM was found overall to be satisfactory, reaching >98% accuracy in most studies. In comparison with classic machine learning approaches, proposed hybrid DL was found to achieve better performance in almost all studies that reported such comparison outcomes. Furthermore, prevention is better than cure. Additionally, employed fuzzy based inference techniques to enhance the prevention procedure using suggestion table.