Intelligent Diagnosis of Heart Diseases Based on Electrocardiographic Signal

Abstract

Background & Objectives: Cardiovascular disease is a leading cause of death worldwide. ECG signals are used to diagnose it. This study aims to eliminate signal noise by converting available wavelets and extracting existing waves. The location-related properties and amplitude of these waves will be extracted to develop a model based on the random forest algorithm for training and evaluating the algorithm. Materials & Methods: This study uses the MIT-BIH dataset, which contains digital ECG signals extracted from Holter bands for different patients at Arrhythmia Hospital from 1975 to 1979. The study applies signal processing and machine learning techniques to classify ECG signals and identify heart patients. The MATLAB software implemented the algorithm, which was evaluated based on accuracy, error rate, TP, FP, Precision, Recall, F-Measure, and ROC criteria. These criteria were determined by a confusion matrix. Results: The study results and comparisons demonstrate that the proposed method is highly effective in detecting heart patients. The proposed method's accuracy was found to be 99%, which is higher than other machine learning methods. Conclusion: The proposed method achieved an accuracy of 99.1957%, surpassing other machine learning methods like support vector machine, neural network, and Bayes.

Keywords

• heart disease
• mit-bih dataset
• random forest algorithm
• wavelet transform