Content-based Re-ranking Method for Recommendation
Xin Huang, Jiameng Bai, Peng Jia, Min Li · IOP Conference Series Materials Science and Engineering · 2020
Abstract The recommendation system is one of the effective tools to solve information overload. Most of the current deep learning recommendation algorithms only focus on the accuracy of the recommendation results by learning users’ preferences. However, the diversity of the recommendation results is neglected, resulting in homogeneous recommendation results, which reduces users’ satisfaction. Diversity can not only effectively avoid over-fitting, but also comprehensively consider multiple dimensions to improve the quality of recommendation results. Therefore, we propose a content-based re-ranking method (CBR) for recommendation systems. The proposed method can make full use of the data set, and effectively supplement and re-rank the recommended results produced by some deep learning algorithms based on the edge information such as tags to be recommended from different perspectives, thus can effectively improve the diversity while preserving the accuracy of recommendation results. Experimental results demonstrate the significant improvements of the proposed re-ranking method.