The Research Of A Data Mining Model Based On Fuzzy Neural Network

Data Mining is a high process of discovering reasonable, novel, valid, understandable information from large amount of data stored in database, data warehouse or other information system. It is an intersecting course involving database, artificial intelligence, statistics, parallel proceeding and so on.Both neural network and fuzzy logic are valid method for data mining. Neural network is good at learning, but the result is not easy to understand. The rule based on reason method of fuzzy logic is easy to comprehend, but it is difficult to get the fuzzy rules for lacking the ability for learning. Fuzzy neural network (FNN) gathered all the merits of neural network and fuzzy logic. It not only can handle with fuzzy rules, but also have the ability to learn. The power to handle nonlinear, fuzzy information is extremely outstanding.This paper introduces some basic theories, model and methods of fuzzy logic, neural network, and FNN. Then turn to Adaptive Network-based Fuzzy Inference Systems (ANFIS), introduce the structure, learning method of ANFIS in detail. Referring to the theory of data mining, it introduces the common method and step to building a FNN based data mining model. Then it builds a stock prediction model and a cloth type classified model in this way. Compared the old, traditional methods, it is more precious. It proved that it is a valid, advanced method to "build a data mining model.