Using nonlinear features and logistic regression for epilepsy detection with linear complexity

In this specific research study, a remarkably accurate and significantly simplified approach has been presented.

This research is encompasses three crucial stages. Firstly, the length of the signal is effectively diminished to an optimal magnitude through the utilization of a technique widely known as windowing. This technique plays a pivotal role in reducing the signal to an ideal size, ensuring the subsequent stages are executed with utmost precision. Secondly, the pertinent features are extracted from the shortened signal, specifically focusing on the Fractal Dimension, the Hurst Exponent, and the Ratio of Determinism to Recurrence Rate. These features are chosen due to the inherent nonlinear nature of the signal, as they provide valuable insights into the complex patterns and structures present within. Lastly, the extracted features undergo Logistic Regression, a widely employed classification algorithm, to effectively categorize and classify them. This step plays a crucial role in providing a clear and concise understanding of the underlying characteristics of the signal.

The implementation of the proposed method not only achieves an outstanding accuracy rate of 99.66%, but it also exhibits a linear time complexity, ensuring efficient processing. Additionally, this method leads to a significant reduction in the length of EEG signals, which is of utmost importance in practical applications.

The primary objective of this proposed method revolves around the introduction of an online approach that can seamlessly integrate into healthcare systems. This objective is derived from a comprehensive analysis and evaluation of the obtained results, ensuring the method's practicality and effectiveness are thoroughly assessed.