Improving the Diagnosis of Arrhythmia using a Combination of Neural Networks in a Hierarchical Way

Heart diseases are one of the most common types of diseases, which cause the death of many people. Arrhythmias are an irregular heartbeat that causes the heart to beat abnormally fast (tachycardia) or slow (bradycardia). Therefore, the identification and classification of cardiac arrhythmias using ECG signals is of great importance. This research aimed to provide a data mining-based model to improve the diagnosis of previous arrhythmia.

In this descriptive-analytical study, the UCI reference dataset, which consists of 452 samples with 279 features, was used. The samples were categorized into five classes for the detection and identification of different types of cardiac arrhythmias. The algorithm employed in this research is a combination of hierarchical neural networks (expert system combination).

In all networks, 70% of the samples were used for training, while the remaining 30% were used for testing. After modeling and comparing the generated models and recording the results, the prediction accuracy for cardiac arrhythmia in the absence of combination learning reached 89.5%, and it increased to 93.5% after employing the hierarchical expert combination approach.

The results of this research show that the proposed method based on the combination of neural networks in a hierarchical form, which leads to the specialization of the task of each class, can have better performance compared to similar models in diagnosing cardiac arrhythmia.