Data Augmentation with GAN increases the Performance of Arrhythmia Classification for an Unbalanced Dataset
This addresses the problem of unbalanced datasets for researchers in medical AI, but it is incremental as it applies an existing method to a specific domain.
The study tackled the data shortage problem in arrhythmia classification by generating synthetic ECG signals using GANs, which increased the performance of the machine learning system.
Due to the data shortage problem, which is one of the major problems in the field of machine learning, the accuracy level of many applications remains well below the expected. It prevents researchers from producing new artificial intelligence-based systems with the available data. This problem can be solved by generating new synthetic data with augmentation methods. In this study, new ECG signals are produced using MIT-BIH Arrhythmia Database by using Generative Adversarial Neural Networks (GAN), which is a modern data augmentation method. These generated data are used for training a machine learning system and real ECG data for testing it. The obtained results show that this way the performance of the machine learning system is increased.