Recommended Based On User Character Collaborative Filtering

With the expanding of information on the Internet, traditional search engines have been difficult to meet the users’personalized needs to access information. To solve this problem, in many areas, the recommendation system is being more widely used. Collaborative filtering recommendation is the most widely used and most successful one of the recommended strategies. However, there is also a cold start, sparsity and scalability issues.Currently, many scholars have study these three issues, but the analysis emphasis on data but ignores the users. It analysis user similarity from data perspective but lack of depth analysis of the users. There is few filtering recommendation strategy take user character into account.For the current study the existence of the problem, the paper studied the utilize of the user character in collaborative filtering recommendation. The paper proposed collaborative filtering method based on user character, character traits is pointed out that the user to calculate the similarity between users, and accordingly select the user nearest neighbor, this approach can solve the sparsity problem. Meanwhile, the paper also presents a comprehensive weighted considering the traditional user ratings and similarity calculation method based on the user’s personality traits. The paper also pointed out that the project can be based on characteristics of the project to be divided, and the combination of user character build character-Item preference matrix. Then the recommended to the user or project can be based on the matrix, the method can solve the cold start problem.Finally, the paper recommended strategy proposed by the experiment. By designing the questionnaire, collected user character and the film’s score. The experimental results show that under the same conditions, the character-based user collaborative filtering recommendation algorithm with high precision than traditional collaborative filtering algorithm, and the two weighted consider recommendation algorithm has higher accuracy.