As-index based Collaborative Filtering recommendation algorithm

Most recommendation systems employ collaborative filtering (CF) for formulating suggestions of items relevant to users´ interests, which commonly uses k-nearest neighbors searching algorithm (kNN) and recommends an item to a user based on the users´ rating table. With the users´ number increasing, it meets the real-time problem and the scalable problem. In this paper, we propose the original AS-INDEX based CF (ASCF). ASCF firstly proposes the index structure (AS-INDEX) based on angular similarity, which refers to the axis and a reference-line to organize the rating table into some shell-hyper-cones, and linearly stores them. Then it determines the storage location for the active user, making a hyper-cone which takes the line connecting the origin point and the user vector as the axis, and searches the hyper-cone for k-nearest neighbors of the user to do the recommendation. ASCF can improve the performance and solve those current shortcomings. We finally demonstrate that our method outperforms the existing methods through experiments using the Jester´s dataset.