Deep Neural Network with Hyperparameter Tuning for Detection of Heart Disease

Citations Over TimeTop 10% of 2021 papers

Abstract

Heart disease causes the most deaths in the world with around 17.89 million people dying each year. Detecting heart disease at an early stage is needed so that further action can be done on the patient. Many researchers have conducted studies about computer-assisted diagnosis system for heart disease. This research presents a heart disease detection method using a deep neural network with hyperparameter tuning. Hyperparameter tuning is done using grid search, random search, and Bayesian optimization. In terms of tuning time, random search spends less time than Bayesian optimization and grid search. In terms of classification performance results, Bayesian optimization produces higher accuracy than grid search and random search. The classification performance of DNN with Bayesian optimization on the testing resulted in an accuracy of 91.67%, a sensitivity of 95.83%, a specificity of 88.89%, a precision of 85.19%, an F1-score of 90.20%, and an AUC value of 0.9514. It indicates that DNN with Bayesian optimization is preferable to be used in detecting heart disease.

Related Papers

• → Initializing Bayesian Hyperparameter Optimization via Meta-Learning(2015)394 cited
• → Deep Neural Network with Hyperparameter Tuning for Detection of Heart Disease(2021)23 cited
• → Bayesian Optimization for Optimizing Retrieval Systems(2018)22 cited
• → Evaluation of Hyperparameter-Optimization Approaches in an Industrial Federated Learning System(2022)14 cited