Adaptive Network Fuzzy Inference Systems for Classification in a Brain Computer Interface
Vahid Asadpour, Mohammd Reza, Reza Fazel-Rezai · InTech eBooks · 2013
The first kind of FIS is designed based on the ability of fuzzy logic to model human percep‐ tion. These FIS elaborates fuzzy rules originates from expert knowledge and they are called fuzzy expert system. Expert knowledge was also used prior to FIS to construct expert systems for simulation purposes. These expert systems were based on Boolean algebra and were not well defined to adapt to regressive intrinsic of underlying process phenomena. Despite that, fuzzy logic allows the rules to be gradually introduced into expert simulators due to input in a knowledge based manner. It also depicts the limitations of human knowledge, particular‐ ly the ambiguities in formalizing interactions in complex processes. This type of FIS offers high semantic degree and good generalization ability. Unfortunately, the complexity of large systems may lead to high ambiguities and insufficient accuracies which lead to poor performances [1].