An hybrid intelligent computational modular with back-propagation network

Back-propagation neural network model (BPNN) is an intelligent computational model based on stylebook learning. This model is different from the traditional adaptability symbolic logic reasoning method based on knowledge and rules. At the same time, BPNN model has shortcomings such as: the slowly convergence speed and partial minimum. In the process of adaptability evaluation, the factors were diverse, complicated and uncertain, so an effectual model should adopt the technique of data mining method and fuzzy logic technologies. In this paper, the author ameliorated the back-propagation of BPNN and applied the fuzzy logical theory for dynamic inference of fuzzy rules. Authors also give detailed description on training and experiment process of the novel model.