Evaluation of performance of Cloud based Neural Network models on Arrhythmia Classification

Arrhythmia classification is always a subject of keen interest in medical sciences as it aids the diagnostic process. Cloud-based real-time cardiac monitoring models are emerging in the market. These monitoring models can compute very intensive tasks in real time and have found a lot of application in Medical Diagnostics. Several cloud-based methods have been proposed, and its total functionality is evaluated. In this paper, we propose an evaluation of different neural network models. The signal is transformed into the wavelet domain, and noise removal is carried out by wavelet de-noising post filtering. The features are extracted from the processed signal and are transmitted to the cloud where predictive models are applied to the extracted features to predict the class of arrhythmia thus aiding the medical diagnostic process.