Dimension reduction in EEG data using Particle Swarm Optimization

摘要: EEG data contains high-dimensional that requires considerable computational power for distinguishing different classes. Dimension reduction is commonly used to reduces the necessary training time of classifiers with some degree accuracy lost. The dimension usually performed on either feature or electrode space. In this study, a new method reduce number electrodes and features using variations Particle Swarm Optimization (PSO) used. variation in terms parameter adjustment adding mutation operator PSO. results are assessed based percentage, potential selected performance An Extreme Learning Machine (ELM) as primary classifier evaluate sets by Two alternative such Polynomial SVM Perceptron further evaluation reduced data. indicate PSO reducing up 99% minimal