Comparison of the Performances of Selected EEG Electrodes with Optimization Algorithms in P300 Based Speller Systems
MetadataShow full item record
Nowadays, the evoked potentials obtained from EEG signals are frequently used in Brain Computer Interfaces (BCI) in order to ensure that individuals with complete paralysis are communicated with the environment. However, EEG signals have a multi-channel structure due to the nature, and this situation increases the data size. Therefore, the working costs of BCIs are increasing. At this stage, reduce the number of electrodes is a solution to reduce the process speed and load in BCI systems. Different techniques are used for this purpose. Some of these are the optimization methods produced by the movement of living beings in nature. In this study, the effects of reduced channels by using Bee Algorithm (BA), Particle Swarm Algorithm (PSO), Differential Evolution (DE), Genetic Algorithm (GA) used in P300 hyphenation systems to reduce the number of electrodes were investigated. As a result of the study, between the accuracy rates obtained with 64 channels and the accuracy rates obtained with channel selecting, (the LS-SVM classifier for User B by B-BA) is seen an about 20% increase. Optimization algorithms are a suitable method to reduce the cost of system operation.