Certain Investigation on Performance of Decomposition Algorithms in EEG-based Sleep Data Analysis

Sleep-related diseases have become a growing concern in contemporary society due to their profound impact on health and well-being. The study of sleep disorders requires a comprehensive understanding of the underlying physiological mechanisms, with EEG signals from the brain serving as a crucial diagnostic tool. This research delves into exploring various decomposition algorithms applied to EEG signals, aiming to ascertain the most effective approach for identifying and diagnosing sleep-related issues. Decomposition plays a huge role in signal analysis such as EEG, by dividing the signals into various forms for further study into that data. Choosing an efficient decomposition algorithm for the dataset determines the quality of the study and can be used to analyze the sleep pattern. Further, the decomposed signals can be used to extract meaningful insights and can be fed into deep learning algorithms for classification purposes. When applying a deep learning algorithm, Wavelet Packet Decomposition (WPD) gives an accuracy of 99.95%, which leads to the conclusion regarding efficiency and performance.