The Research Of Individuation Recommendation Algorithm Based On Item Classific Comparability And User's Multiplex Interesting

With the popularization of Internet and the development of E-Commerce, the structure of E-Commerce web site became more and more complex. This situation made it hard for consumers to find the products and services they wanted. To address this issue, recommendation systems were proposed to suggest products and to provide consumers with information to help them decide which products to purchase. Recommendation systems can enhance 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 E-Commerce IT technologies, more and more research papers about recommendation systems in E-Commerce appeared in many kinds of conferences and journals.On the aspect of theory research and factual application, E-commerce recommender systems have developed rapidly. Recommendation approaches are in the heart of recommender system, so which approach is adopted is crucial to the success of recommending quality as well as efficiency. Today, five approaches are used by recommender systems to generate recommendations, namely knowledge engineering, collaborative filtering, content-based, hybrid and data mine approaches. Collaborative filtering has been very successful in both research and practice. But,with expansion of E-commerce system's size, collaborative filtering approach suffer from many challenges, for instance, quality of recommendations, scalability, sparsity,cold-start problem. The main research works in the thesis included research of recommendation quality, research of real-time requirement of recommendation.The main research results of this thesis are as follows:1. The trational comparability measurement method didn't consider the influence of item' s classificatory attribute. So the result was inadequate precision. Contraposing to this problem, this paper put forward a new method about the item similarity, first, Fuzzy clustering technology is used for user. This method can make single user ratings become user colony rating to item, this can construct dense rating matrix of user colony-item. We use both classificatory of item and user rating to calculate the item's comparability. In order to improving the nicety.2. A new method about finding user's neighbors is proposed: first the Item-Based collaborative filtering recommendation algorithm is used to predit the rating for the item which is not rated. this can low the data sparisty, because of the user mutiplex interesting was existed, so when the finally predicted grade of the target item was calculated, we look for the user's neighbors don't need to calculate in the whole item space, but in the neighbor's set space of target item, this method can improve the veracity of the predicted grade.