Implementation and Analysis of an Optimized Feature Selection Algorithm for Stress Detection

Feature selection is crucial for optimizing machine learning in high-dimensional data domains like biomedical signal processing. This paper introduces a novel methodology that enhances feature selection for stress detection by using an electroencephalography (EEG) signal, integrating the Archimedes optimization algorithm (AOA) with the Analytical Hierarchical Process (AHP). AOA optimizes biosignal feature selection by balancing exploration and exploitation phases to refine significant features crucial for stress detection. Meanwhile, AHP systematically evaluates biosignal criteria to prioritize features for stress detection. The proposed methodology employs AOA-AHP integration to identify relevant biosignal features efficiently. Experimental results demonstrate significant improvements in classification performance, achieving optimal accuracy. Additionally, the methodology ensures computational efficiency by reducing trainable parameters and training time. The robust performance of the AOA-AHP framework across various feature set sizes underscores its suitability for real-world applications in healthcare diagnostic and monitoring systems.