The Research And Design Of A Collaborative Filtering Recommendation System Based On The User And The Weight Of Resources

With the popularization of Internet and the development of Services information, the System of Services information can provide users with more and more choices and at the same time its structure has becoming more and more complex. This situation made it hard for consumers to find the information and services they wanted. To address this issue, recommendation systems were proposed to provide the target user with information they interested in based on the existing information, so users can be more convenient to find the information they need. With the development of commercial information, the application of recommendation systems in the Service System has become more and more widely. So to provide an accurate and real-time recommendation system is the main goal to accelerate the industrialization of today's information.Recommender systems make information filtering for user by predicting user's preference to items. They apply knowledge discovery techniques to the problem of making personalized recommendations. Collaborative filtering is becoming a popular technique for reducing information overload. However, most of current collaborative filtering algorithms have several major limitations: accuracy, data sparsely and cold start problem. There are many algorithms that combining collaborative filtering and content-based information filtering method have been proposed to solve the problems. However, these algorithms can not show the user's interest very well, so the result of recommendation can not meet the needs of users in most of time, and in many cases, the interest of user will be changed with the over time. In order to improve the quality of the recommendation, some of characteristics of users, as well as the corresponding changes must be taken into account; In addition, the values of different items will be different for the users, and these values can be dug form the user's score information. Therefore, the value of the resource itself can also become an important factor in recommendation systems.How to improve the quality of Recommendation is the goal of this research, because of the accuracy of the recommend quality depends on the selection of neighbor users to a great extent. So it must to improve the accuracy of similar user to improve the quality of recommendation, at the same time, it also the emphases of the research. This study is based on the relatively complete user's information, combine the weight of resources with some characteristics of user's, offer an recommendation method based on user and the weight of resource, and this method make a corresponding improvement compare with the traditional recommendation methods.The main researches of this paper are mainly reflected in the following areas:1) According to the table of user-item score, set up the corresponding vector space model, and this paper is mainly include the model of the characteristics of user and the items.2) Compare with the traditional collaborative filtering technology, the time function and the impact of similarity between users, which changes with the change of user's interest, are taken into account in my research. In additional, according to the value of items, an algorithm based on the characteristics of user and item has been designed. We can use the algorithm to calculate the similarity between users to form the similar user set.3) On the basis of the user's information in the similar user set, we can use the advanced collaborative filtering algorithm to predict the user's score to the items that have not been scored. and then show the user specified number of items by the order of high to low according to the predicting results.The main innovation of this paper is combining the characteristics of user with the weight of resources based on the traditional collaborative filtering algorithm, improving the traditional algorithm so that it can better adapt to the changes of user's interest which will be changed with the time. m this paper, we also take the weight of resources into account so that it can improve the finally prediction, and enhance the accuracy and timeliness of the recommendation to some extent.