Sentiment detection in social networks and in collaborative learning environments

作者：

Highlights：

• A Mixed Graph of Terms (mGT) can be used effectively for sentiment detection.

• The discriminative power of the Mixed Graph of Terms structure is improved by the use of annotated lexicon.

• Sentiment analysis performed through mGT shows encouraging results on social nets’ posts as well as standard datasets.

摘要

•A Mixed Graph of Terms (mGT) can be used effectively for sentiment detection.•The discriminative power of the Mixed Graph of Terms structure is improved by the use of annotated lexicon.•Sentiment analysis performed through mGT shows encouraging results on social nets’ posts as well as standard datasets.

论文关键词：Information extraction,Sentiment analysis,Latent Dirichlet allocation

论文评审过程：Available online 27 December 2014, Version of Record 29 July 2015.

论文官网地址：https://doi.org/10.1016/j.chb.2014.11.090