Hybrid self organizing map and probabilistic quadratic loss multi-class support vector machine for mental tasks classification

摘要: Brain computer interface provides communication opportunity between the brain and environment around a person with severe motor disabilities. However, implementation of such interfaces requires good signal processing scheme, whose performances depend principally on technique used to select best features, classification perform discrimination different categories. This work proposes new hybrid structure based two stages supervised unsupervised learning. The first stage consists Self Organizing Map which allows cluster redundant irrelevant features descriptors. second uses, Probabilistic Quadratic Loss Multi-Class Support Vector Machine for final classification. Five mental tasks namely: Baseline, Multiplication, Letter, Rotation, Counting are considered, an average accuracy 81.73% 91.90% is achieved. result shows effectiveness proposed method enhance performance electroencephalogram problem.