Rating prediction using review texts with underlying sentiments
详细信息 查看全文
- 作者：Dongjin Yu ; yudj@hdu.edu.cn" class="auth_mail" title="E-mail the corresponding author ; Yunlei Mu muyunlei@gmail.com" class="auth_mail" title="E-mail the corresponding author ; Yike Jin jyk0406@126.com" class="auth_mail" title="E-mail the corresponding author
- 关键词：Collaborative filtering ; Latent factor model ; Information retrieval ; Cold-start ; Sentiments
- 刊名：Information Processing Letters
- 出版年：2017
- 出版时间：January 2017
- 年：2017
- 卷：117
- 期：Complete
- 页码：10-18
- 全文大小：1117 K
文摘
- class="listitem" id="list_ls0010">
- class="label">•
Transformation structured to link user/item avg. rating with sentiment probability.
- class="label">•
A model is proposed to combine HFT with ASUM based on the transformation.
- class="label">•
Latent factors are linked with topic probabilities to model aspects of sentiments.
- class="label">•
An optimization algorithm is given to optimize the parameters of RAS alternately.
- class="label">•
Prediction accuracy and convergence rate are demonstrated on real-world datasets.