Semi-supervised learning with generative model for sentiment classification of stock messages

作者：

Highlights：

• The psychology emotion model is employed into the process of message generation.

• Words are divided into three types with different relevant degrees to emotions.

• Semi-supervised learning is realized by GEM-CW to classify stock message sentiment.

摘要

•The psychology emotion model is employed into the process of message generation.•Words are divided into three types with different relevant degrees to emotions.•Semi-supervised learning is realized by GEM-CW to classify stock message sentiment.

论文关键词：Sentiment analysis,Generative model,Semi-supervised learning,Stock message board

论文评审过程：Received 6 February 2019, Revised 7 February 2020, Accepted 7 May 2020, Available online 11 May 2020, Version of Record 27 May 2020.

论文官网地址：https://doi.org/10.1016/j.eswa.2020.113540