Collaborative Filtering Based on Improved Fuzzy C-means Algorithm

Abstract

This paper studies algorithms for recommending suitable companies for college graduates. Firstly, the SimRank algorithm is used to calculate the students' similarity matrix.Secondly the students are initially clustered using canopy, the cluster center is retained, and the final clustering result is obtained by fuzzy c-means algorith(FCM). Finally the results are sorted to obtain the final recommendation result. Through the comparative experiment, the algorithm route proposed in this paper is compared with the following four situations:Use the unimproved SimRank algorithm, and other technical routes remain unchanged.Use the improved SimRank algorithm, and use the Canopy algorithm alone.Use the improved SimRank algorithm, and ues the fuzzy c-means algorithm alone. Use the improved SimRank algorithm, use the k-means clustering algorithm alone. In the end, it was found that the algorithmic route proposed in this paper has better accuracy and recommended recall index. In the end, it was found that the algorithmic route proposed in this paper has better accuracy and recommended recall index.