Collaborative Filtering Recommendation Based on User Character
Zhi Chao Quan · Applied Mechanics and Materials · 2013
Traditional collaborative filtering recommendation system put emphasis on data but ignores the users. The recommendation of the similarity analysis emphasizing too much from the data perspective lacks depth analysis of the users without regarding the similarity from the users perspective. In this paper, user character is introduced to improve the user model, and two character-based collaborative filtering recommendations will be proposed: one is to compute user similarity from the user character perspective and select nearest neighbor, and then generates recommendation; another is based on the character-item rating matrix, and then make recommendation to the prospective users. These two ideas can well make up for the inadequacies of the current collaborative filtering recommendation system.