QPLSA: Utilizing quad-tuples for aspect identification and rating

作者：

Highlights：

• We propose two novel aspect rating prediction approaches, i.e., Quad-tuple prediction and Expectation prediction.

• We analyze and investigate the performance of the proposed aspect rating prediction methods in contrast with Local Prediction and Global Prediction.

• We experimentally inspect the influence of aspect rating variance for different rating prediction approaches.

摘要

•We propose two novel aspect rating prediction approaches, i.e., Quad-tuple prediction and Expectation prediction.•We analyze and investigate the performance of the proposed aspect rating prediction methods in contrast with Local Prediction and Global Prediction.•We experimentally inspect the influence of aspect rating variance for different rating prediction approaches.

论文关键词：Quad-tuple PLSA,Aspect mining,Sentiment analysis

论文评审过程：Received 22 October 2012, Revised 15 May 2014, Accepted 18 August 2014, Available online 17 September 2014.

论文官网地址：https://doi.org/10.1016/j.ipm.2014.08.004