Study Of The Cold Start Problem In Recommender Systems Based On Classification Method

The Recommender We analyzed and summarized the existing collaborative filtering cold start problem solving method, and divides them into three categories, respectively determine the neighbor similarity in collaborative filtering, similarity calculation and prediction score three stages, the traditional method is improved.System (RS) is an efficient tool for decision makers that assists in the selection of appropriate items according to their preferences and interests. This system has been applied to various domains to personalize applications by recommending items such as books, movies, songs, restaurants, news articles and jokes, among others. An important issue for the RS that has greatly captured the attention of researchers is the new user cold-start problem,which is related to recommendations for novel users or new items. In case of new users, the system does not have information about their preferences in order to make recommendations.In this thesis, we analyzed and summarized the existing collaborative filtering cold start problem solving method, and divided them into three categories, respectively determine the neighbor similarity in collaborative filtering, similarity calculation and prediction score three stages, the traditional method is improved.we propose a model for the solution of the cold start problem, get the product prediction score from a new user through three steps, the neighbors of the new users are found out by considering the demographic data and similarity based technology,.The new user has similar characteristics is definded as a neighbor. This idea is people with similar background and characteristics are more likely to have similar hobbies. In this step, users are separated by known classifier. Then, each new user is divided into a group, based on a score prediction mechanism of new user of the product evaluation. Calculate the final score using a weighting model, so that developers can focus on a particular property or choose a more equitable way.Finally, the performance of the proposed technique,which provided by the Grouplens study group data set, is shown by experiment. This technology has a large number of registered users in the system who perform better than others, in this case, the system has smaller MAE, thereby increasing the rating prediction accuracy.