An efficient method for classifying motor imagery using CPSO-trained ANFIS prediction

Abstract

Adaptive neuro-fuzzy inference system (ANFIS) is a hybrid of two soft computing methods of the artificial neural network (ANN) and fuzzy logic. Fuzzy logic has the advantage to change the qualitative aspects of human knowledge and insights into the process of precise quantitative analysis. However, it does not have a defined method that can be used as a guide in the process of transformation and human thought into rule-based fuzzy inference system (FIS). The fuzzy system cannot learn or adapt itself to the new environment, while the ANN is ambiguous to the user. By combining these two methods, the ANN becomes more transparent, and the fuzzy system takes on the ability of learning. With this combination, a more effective model in the medical domain could be built. In this paper, the ANFIS trained with corrected particle swarm optimization (CPSO), is proposed to classify the brain–computer interface (BCI) motor imagery mental tasks based on electroencephalography signals. The dataset is used in this paper, is BCI competition IV dataset. For evaluating the proposed method in order to obtain more classification rate, it is compared to the conventional ANFIS and the ANFIS trained by other evolutionary algorithms (EAs) such as genetic algorithm, particle swarm optimization, differential evolution, and biogeography-based optimization (BBO) that are more closely to the CPSO algorithm than the other EAs. The results showed that the ANFIS prediction trained by CPSO algorithm has more performance compared to conventional ANFIS prediction and ANFIS trained by other EAs.