Lexicon-Enhanced LSTM With Attention for General Sentiment Analysis
Long short-term memory networks (LSTMs) have gained good performance in sentiment analysis tasks. The general method is to use LSTMs to combine word embeddings for text representation. However, word embeddings carry more semantic information rather than sentiment information. Only using word embeddi...
Main Authors: | , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
IEEE
2018-01-01
|
Series: | IEEE Access |
Subjects: | |
Online Access: | https://ieeexplore.ieee.org/document/8513826/ |
id |
doaj-9016b5090dae45578fe2efb4ba790e6a |
---|---|
record_format |
Article |
spelling |
doaj-9016b5090dae45578fe2efb4ba790e6a2021-03-29T21:34:40ZengIEEEIEEE Access2169-35362018-01-016718847189110.1109/ACCESS.2018.28784258513826Lexicon-Enhanced LSTM With Attention for General Sentiment AnalysisXianghua Fu0https://orcid.org/0000-0003-4431-3386Jingying Yang1Jianqiang Li2Min Fang3Huihui Wang4Faculty of Arts and Sciences, Shenzhen Technology University, Shenzhen, ChinaCollege of Computer Science and Software Engineering, Shenzhen University, Shenzhen, ChinaCollege of Computer Science and Software Engineering, Shenzhen University, Shenzhen, ChinaExperimental and Creative Practice Education Center, Harbin Institute of Technology, Shenzhen, ChinaDepartment of Engineering, Jacksonville University, Jacksonbville, FL, USALong short-term memory networks (LSTMs) have gained good performance in sentiment analysis tasks. The general method is to use LSTMs to combine word embeddings for text representation. However, word embeddings carry more semantic information rather than sentiment information. Only using word embeddings to represent words is inaccurate in sentiment analysis tasks. To solve the problem, we propose a lexicon-enhanced LSTM model. The model first uses sentiment lexicon as an extra information pre-training a word sentiment classifier and then get the sentiment embeddings of words including the words not in the lexicon. Combining the sentiment embedding and its word embedding can make word representation more accurate. Furthermore, we define a new method to find the attention vector in general sentiment analysis without a target that can improve the LSTM ability in capturing global sentiment information. The results of experiments on English and Chinese datasets show that our models have comparative or better results than the existing models.https://ieeexplore.ieee.org/document/8513826/Sentiment lexiconsentiment embeddingword embeddingattention vectorsentiment analysis |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Xianghua Fu Jingying Yang Jianqiang Li Min Fang Huihui Wang |
spellingShingle |
Xianghua Fu Jingying Yang Jianqiang Li Min Fang Huihui Wang Lexicon-Enhanced LSTM With Attention for General Sentiment Analysis IEEE Access Sentiment lexicon sentiment embedding word embedding attention vector sentiment analysis |
author_facet |
Xianghua Fu Jingying Yang Jianqiang Li Min Fang Huihui Wang |
author_sort |
Xianghua Fu |
title |
Lexicon-Enhanced LSTM With Attention for General Sentiment Analysis |
title_short |
Lexicon-Enhanced LSTM With Attention for General Sentiment Analysis |
title_full |
Lexicon-Enhanced LSTM With Attention for General Sentiment Analysis |
title_fullStr |
Lexicon-Enhanced LSTM With Attention for General Sentiment Analysis |
title_full_unstemmed |
Lexicon-Enhanced LSTM With Attention for General Sentiment Analysis |
title_sort |
lexicon-enhanced lstm with attention for general sentiment analysis |
publisher |
IEEE |
series |
IEEE Access |
issn |
2169-3536 |
publishDate |
2018-01-01 |
description |
Long short-term memory networks (LSTMs) have gained good performance in sentiment analysis tasks. The general method is to use LSTMs to combine word embeddings for text representation. However, word embeddings carry more semantic information rather than sentiment information. Only using word embeddings to represent words is inaccurate in sentiment analysis tasks. To solve the problem, we propose a lexicon-enhanced LSTM model. The model first uses sentiment lexicon as an extra information pre-training a word sentiment classifier and then get the sentiment embeddings of words including the words not in the lexicon. Combining the sentiment embedding and its word embedding can make word representation more accurate. Furthermore, we define a new method to find the attention vector in general sentiment analysis without a target that can improve the LSTM ability in capturing global sentiment information. The results of experiments on English and Chinese datasets show that our models have comparative or better results than the existing models. |
topic |
Sentiment lexicon sentiment embedding word embedding attention vector sentiment analysis |
url |
https://ieeexplore.ieee.org/document/8513826/ |
work_keys_str_mv |
AT xianghuafu lexiconenhancedlstmwithattentionforgeneralsentimentanalysis AT jingyingyang lexiconenhancedlstmwithattentionforgeneralsentimentanalysis AT jianqiangli lexiconenhancedlstmwithattentionforgeneralsentimentanalysis AT minfang lexiconenhancedlstmwithattentionforgeneralsentimentanalysis AT huihuiwang lexiconenhancedlstmwithattentionforgeneralsentimentanalysis |
_version_ |
1724192670193025024 |