Affordable Access

Access to the full text

Tag-Extended Collaborative Filtering Recommendation Algorithm

Authors
  • Chen, Hailong1
  • Yan, Wuyue1
  • Sun, Haijiao1
  • Cheng, Miao1
  • 1 Harbin University of Science and Technology, Harbin, 150080, China , Harbin (China)
Type
Published Article
Journal
SN Computer Science
Publisher
Springer Singapore
Publication Date
Sep 14, 2020
Volume
1
Issue
5
Identifiers
DOI: 10.1007/s42979-020-00316-7
Source
Springer Nature
Keywords
License
Yellow

Abstract

For most of recommendation algorithms that use the relationship between tags, users, and items have to face the problem of tag sparse. Aiming at the problem of sparse tag data that caused by the user's subjective arbitrariness when labeling items, this paper proposes a tag extension collaborative filtering recommendation algorithm. We test the tag similarity in terms of user behavior and tag semantics, and use the tag similarity to expand the item-tag vector to reduce the sparseness of the matrix generated by the association between items and tags. Then we calculate the item similarity for item-based collaborative filtering recommendation. The core of collaborative filtering recommendation algorithm is to calculate the similarity of users or resources. However, it is necessary to consider the question that the items labeled by a user are not necessarily what the user likes and interest offset when using tag information to calculate the similarity of items. So for improving the accuracy of the recommendation algorithm, we improve the calculation method of similarity. We integrate user rating information and time information into the calculation of similarity to increase the weight of the item that tagged by the user and has a high rating and to reflect the change of user's preference. Finally, the MovieLens dataset is employed to test different algorithms and results show that the proposed algorithm has a higher accuracy.

Report this publication

Statistics

Seen <100 times