中国邮电高校学报(英文) ›› 2021, Vol. 28 ›› Issue (5): 82-90.doi: 10.19682/j.cnki.1005-8885.2021.0026

• Artificial Intelligence • 上一篇    下一篇

News recommendation based on time factor and word embedding

顾亦然,周鹏,杨海根   

  1. 南京邮电大学
  • 收稿日期:2020-11-25 修回日期:2021-02-26 出版日期:2021-10-31 发布日期:2021-10-29
  • 通讯作者: 顾亦然 E-mail:guyrnj@163.com
  • 基金资助:
    国防基础科研;国防重大工程项目;装备发展部仿真预研课题

News recommendation based on time factor and word embedding

  • Received:2020-11-25 Revised:2021-02-26 Online:2021-10-31 Published:2021-10-29
  • Supported by:
    National Defense Basic Scientific Research program of China

摘要:

Existing algorithms of news recommendations lack in depth analysis of news texts and timeliness. To address these issues, an algorithm for news recommendations based on time factor and word embedding ( TFWE) was proposed to improve the interpretability and precision of news recommendations. First, TFWE used term frequency- inverse document frequency ( TF-IDF ) to extract news feature words and used the bidirectional encoder representations from transformers ( BERT ) pre-training model to convert the feature words into vector representations. By calculating the distance between the vectors, TFWE analyzed the semantic similarity to construct a user interest model. Second, considering the timeliness of news, a method of calculating news popularity by integrating time factors into the similarity calculation was proposed. Finally, TFWE combined the similarity of news content with the similarity of collaborative filtering ( CF) and recommended some news with higher rankings to users. In addition, results of the experiments on real dataset showed that TFWE significantly improved precision, recall, and F1 score compared to the classic hybrid recommendation algorithm.


关键词:

news recommendation, time factor, word embedding, user interest model


Abstract:

Existing algorithms of news recommendations lack in depth analysis of news texts and timeliness. To address these issues, an algorithm for news recommendations based on time factor and word embedding ( TFWE) was proposed to improve the interpretability and precision of news recommendations. First, TFWE used term frequency- inverse document frequency ( TF-IDF ) to extract news feature words and used the bidirectional encoder representations from transformers ( BERT ) pre-training model to convert the feature words into vector representations. By calculating the distance between the vectors, TFWE analyzed the semantic similarity to construct a user interest model. Second, considering the timeliness of news, a method of calculating news popularity by integrating time factors into the similarity calculation was proposed. Finally, TFWE combined the similarity of news content with the similarity of collaborative filtering ( CF) and recommended some news with higher rankings to users. In addition, results of the experiments on real dataset showed that TFWE significantly improved precision, recall, and F1 score compared to the classic hybrid recommendation algorithm.



Key words:

news recommendation, time factor, word embedding, user interest model


中图分类号: