Fuzzy support vector machine analysis in EEG classification

ABSTRACT:Brain Computer Interface (BCI) technology, provides a direct electronic interface between brain and computer. It enables people with movement disabilities meet their main needs. BCI systems have three parts as input, output, and a processing algorithm that maintains a relation between input and output. The algorithm has three parts of preprocessing, feature extraction and classification. In this article after pre processing the signal we used fractal features like Petrosian and Sevcik’s methods to extract features. In classification we used fuzzy support vector machines and compared it with three other classifiers. In final we resulted that fuzzy support vector machines with Petrosian fractal features has the most classification accuracy (82%) than others but its computation time with two fractal features as Petrosian and Sevcik’s features is not the best but LDA (linear Discriminate Analysis) with Petrosian fractal features has the best computation time (0.14s).