The Application In The Comprehensive Evaluation Based On Dynamic Fuzzy Neural Network

With the wide application of fuzzy comprehensive evaluation methods in each field, its importance and superiority is becoming more and more obvious and prominent. Though there are many researchers and actual workers at home and abroad to make the comprehensive evaluation model improve and expand continuously, establishing a more perfect evaluation model has become inevitable using the complementary strengths of neural network and fuzzy technology because of intelligent technology development and the mutual infiltration of different disciplines. In recent years dynamic comprehensive evaluation research has caused wide public concern and become the hot spot and frontier in comprehensive evaluation research. Although dynamic comprehensive evaluation problem has aroused the attention of scholars, its study is quite rare because it is more complex. It has not been formed complete framework system. Its many aspects need to further research and richen.Dynamic comprehensive evaluation problem is a kind of theoretical and practical significance. Its application field includes change index with time and the parameters fluctuation larger. In real life, it involves dynamic evaluation of all problems of dynamic economic, managers' performance evaluation, examination and candidate sorting problems. This paper based on the study of the fuzzy theory and fuzzy neural network model, the effective combination of the fuzzy theory and the neural network, the existing disadvantages of comprehensive evaluation models, based on the existed evaluation model using BP and RBF neural network to establish the new dynamic evaluation model. Its main features are: in the course of study (1) the parameters can adjust, (2) the structure can be established in the course of study automatically; introduced the concept of time, namely evaluation index of different time weight coefficient evaluated object evaluation continuously using the same evaluation system at different moment, evaluation results formed a set of "sequence", and rally this "sequence" to get the final evaluation value through time weight vector.The network applies to teachers' teaching quality evaluation in order to test whether the model is effective or practical.