A Collaborative Filtering Algorithm Based on Mixed Similarity

Yang ZOU, Ying-ding ZHAO

Abstract


The traditional collaborative filtering algorithm ignores the influence of user interest and item popularity in the calculation of similarity, which will lead to inaccurate calculation of similarity in terms of sparse data. This paper proposes a collaborative filtering algorithm which is based on mixed similarity algorithm. Both user interest similarity and item popularity similarity are taken into consideration. Results show that the proposed algorithm significantly reduces the calculation error comparing with the traditional algorithm. At the same time, the problem of data sparsity has been alleviated.

Keywords


User interest, Item popularity factor, Data sparsity.


DOI
10.12783/dtcse/msota2018/27587

Full Text:

PDF

Refbacks

  • There are currently no refbacks.