A clonal selection algorithm for the electro encephalography signals reconstruction

Abstract

This paper describes an adaptation of the Clonal Selection Algorithm for the single objective Big Optimization problem “Big-Opt”. Indeed, the electroencephalography “EEG” signals are very sensitive to noising effects of artifacts caused by undesirable internal and external electric sources. the main purpose of the Big-OPT problem is to rebuild the recorded signals in the goal of removing the artifacts as maximum as possible. To this end, an optimization problem is defined. To solve it, a modified search strategy is studied and adapted for the Clonal Selection Algorithm in order to enhance its convergence abilities on large scale optimization. To test the performance of the proposed method, experiments have been conducted over the Big-OPT EEG datasets. A comparison with recent state of the art approaches is also included. The study exhibits the competitive performance of the proposed Clonal Selection Algorithm.