<![CDATA[Persons with disabilities often face limitations in interacting with their environment, thereby requiring brain-computer interface (BCI) technology capable of translating brain signals into commands. Electroencephalogram (EEG) signals used in BCI are complex, non-stationary, and highly susceptible to noise, making the classification of brain activity patterns challenging. This study addresses the problem of improving the accuracy and efficiency of EEG classification while preserving the dominant information within the signals. The novelty of this research lies in the application of the Deep Evolving Denoising Autoencoder (Devdan) algorithm integrated with a comprehensive preprocessing pipeline, including Notch and Bandpass (7–13 Hz) filtering, windowing, standard scaling, and dimensionality reduction using Principal Component Analysis (PCA). This integration enables Devdan to operate more effectively on cleaner and reduced data. Experimental results demonstrate an average classification accuracy of 96.13%, with a loss of 0.83 and a short training time of 2.05 seconds. Devdan proved to be adaptive to variations among respondents with consistent and stable outcomes. These findings indicate that the proposed approach is effective for real-time EEG classification and has strong potential to support the development of BCI technology for persons with disabilities in communication and device control.]]>
Copyrights © 2026
