Based On Java Technology System And Implement The Recommendations Of The Personalized

Now, with the rapid development of computer information technology, network-increasing amount of information, users want to access needed information has become increasingly difficult, how the user to find their own information, have been as many scholars, experts and Internet users concerned with one of the core issues. Personalized recommendation systems have emerged in this context.Collaborative filtering algorithm is the personalized recommendation system recommended by the main algorithms, collaborative filtering model of the basic idea is similar personalized recommendations between users, first through the establishment of appropriate mathematical models to describe the similarity between users, then this Based on the design of efficient algorithms for the target user to find similar users, and resources of interest to users recommend similar to the target user. Called the "correlation of man by man" (people-to-people correlations), that is almost certain the group of users. But the collaborative filtering algorithm in the case of high volume users, the system always takes a lot of time calculating the user's neighbors to generate a list of users recommended for this situation, we need to study a new recommendation algorithm to improve the recommendation the efficiency of the generated list.Community-based collaborative filtering algorithms, collaborative filtering algorithm is a variant form, first of all divided into different user communities, and assuming "the same community of users will have similar interests, and different community of users will not have neighbors "Therefore, the recommendation in the generated list of users, only users in the same community to find neighbors, so that when looking for neighbors to reduce the scope of the generation of a list of recommendations to improve efficiency.This paper will briefly introduce personalized recommendation system and the system filtering algorithm and then using JAVA technology to achieve community-based collaborative recommendation algorithm, and carried out a systematic analysis, if the system we can find k communities, due to the phased learning , and the user can assume that the community needed to learn the information can be found in these communities, if the k-number of users within the community is basically the same, then the classification of community collaboration in community without the community than the classification algorithm for collaborative filtering algorithm efficiency high-k times, and recommended the same effect.