Segmentation of consumer's purchase behavior based on neural network

The present algorithms for artificial neural network, to a certain extent, have various questions such as computational complexity, low accuracy and narrow scope of application. This paper presents a new algorithm for extracting accurate and comprehensible rules from databases via trained artificial neural network using genetic algorithm. The new algorithm does not depend on the ANN training algorithms; also it does not modify the training results. The genetic algorithm is used to find the optimal values of input attributes (chromosome), Xm, which maximize the output function ϕk of output node k. The function ϕk is nonlinear exponential function. The optimal chromosome is decoded and used to obtain a rule belonging to class k. The good result is achieved by applying the new algorithm to a given database for customers buying MP3.