Sentiment analysis based on improved pre-trained word embeddings

作者：

Highlights：

• Improved Word Vectors (IWV) increases the accuracy of existing pre-trained word embeddings.

• We propose a proper combination of lexicons to improve sentiment classification.

• IWV improved the accuracy of Google's pre-trained Word2Vec to over 2%.

摘要

•Improved Word Vectors (IWV) increases the accuracy of existing pre-trained word embeddings.•We propose a proper combination of lexicons to improve sentiment classification.•IWV improved the accuracy of Google's pre-trained Word2Vec to over 2%.

论文关键词：Sentiment analysis,Deep learning,Word embeddings,Word2Vec,GloVe,Natural language processing

论文评审过程：Received 17 February 2018, Revised 5 August 2018, Accepted 25 August 2018, Available online 19 September 2018, Version of Record 29 September 2018.

论文官网地址：https://doi.org/10.1016/j.eswa.2018.08.044