A Similarity Calculating Approach Simulated from TF-IDF in Collaborative Filtering Recommendation
Qilong Ba, Xiaoyong Li, Zhongying Bai · International Conference on Multimedia Information Networking and Security · 2013
In the paper, we proposed a new approach to calculate the similarity between users in recommendation systems. The approach simulates the term frequency -- inverse document frequency (TF-IDF) statistical approach which is popular in the area of searching technology. Through the approach, we can calculate the similarity between the target user and his every neighbor without considering the detailed ratings of the common items. After getting the similarities, we can predict the ratings of those unrated items and finally make the recommendation. The approach we proposed in the paper can not only improve the precision and the real-time, but also increase the scalability of the recommendation system.