The Research On Intelligent Recommendation System Based Neural Network And Fuzzy Logic

With the popularization of Internet and the development of E-Commerce, the structure of web site has become more and more complex. This situation has made it hard for consumers to find the products and services they needed. To solve this issue effectively, recommendation systems are proposed to help consumer decide which products to purchase.Recommendation systems can increase E-Commerce sales by converting browsers into buyers, increasing cross-sell and building loyalty to prevent user losing. Presently recommendation systems have gradually become an important part in Web Intelligence technologies, more and more researches about recommendation systems have appeared in many conferences and journals.This dissertation explores and researches the application of neural network and fuzzy logic in recommendation systems. The main research works in the dissertation include: the recommendation quality, the real-time requirement of recommendation, the recommendation systems based on web mining and arthitecture of recommendation systems. The main study and contributions in this dissertation are given as follows:1) Modified Fuzzy ART algorithm. This dissertation analyzes the original fuzzy ART algorithm, and a modified Fuzzy ART is presented. The experimental results also show that the new algorithm has better classification precision than the old one.2) An e-commerce consumer classification method based neural network and web mining. Due to the expansion of E-Commerce systems, the magnitudes and scales of users and commodities grow rapidly. This situation makes the real-time requirement of recommendation system be difficult to be satisfied and the quality of recommendation systems decreases dramatically. To address and solve this issue effectively, we propose a consumer classification approach based on neural network and web mining. Our approach collects the consumer history imformation to construct user model, at the same time neural network is emplyed to classify user model. And the classification results of user model represent the consumer clusters. The experiemental results have illustrated that this approach can efficiently improve the real-time response speed of recommendation systems.3) An architecture of recommendation systems of consumer electronic products based on fuzzy logic. As there are a great number of products provided by e-commerce...