Arrhythmia Classification Using Fractal Dimensions and Neural Networks
Ben Ali Sabrine, Taoufik Aguili · Advances in intelligent systems research/Advances in Intelligent Systems Research · 2022
According to statistics, there has been a big increment in death in consequence of failures worldwide.Electrocardiogram was chosen as a possible implement for diagnosing cardiovascular diseases, it is a test that records the electrical activity given by the heart muscle and how it contracts.In this vein, our work is reported to analyze this low-cost and widely available signal.One of major issues that arise during the analysis of the electrical activity in the heart is noise reduction in electrocardiogram signals.The best bothersome noise sources have frequency components within the electrocardiogram spectrum.Thus, noises are difficult to take away using standard filtering procedures.Indeed, we show how wavelets can be used to denoise such signals.For this reason, electrocardiogram signal is considered as a selfsimilar object.As a result, fractal analysis can be used to make better use of the information gathered.The fractal dimension is considered the best explanation of the electrocardiogram signal that can account for its hidden complexity.This paper uses the fractal dimension to introduce a new technique for the simple classification of arrhythmias from electrocardiogram signals.We used neural networks to improve our classification results, as variety is one of the most active research and application areas for neural network