Paper Info
Title
Polarity Classification Of Tweets Considering The Poster'S Emotional Change By A Combination Of Naive Bayes And Lstm
Abstract
Twitter, as a popular social networking service, is used all over the world, with which users post tweets for various purposes. When users post tweets, an emotion may be behind the messages. As the emotion changes over time, we should better consider their emotional changes and states when analyzing the tweets. In this study, we improve polarity classification by considering the poster's emotional state. Firstly, we analyze the sentence structure of a tweet and calculate emotion scores for each category by Naive Bayes. Then, the poster's emotion state is estimated by the emotion scores, and a prediction model of emotional state is created by Long Short Term Memory (LSTM). Based on the predicted emotional state, weights are added to the scores. Finally, polarity classification is performed based on the weighted emotion scores for each category. In our experiments, our approach showed better accuracy than other related studies.
Year
DOI
Venue
2019
10.1007/978-3-030-24289-3_43
COMPUTATIONAL SCIENCE AND ITS APPLICATIONS - ICCSA 2019, PT I: 19TH INTERNATIONAL CONFERENCE, SAINT PETERSBURG, RUSSIA, JULY 1-4, 2019, PROCEEDINGS, PT I
Keywords
Field
DocType
Twitter, Polarity classification, Naive Bayes, Deep learning
Social network,Naive Bayes classifier,Computer science,Long short term memory,Computer network,Artificial intelligence,Natural language processing,Emotional Changes,Deep learning,Sentence
Conference
Volume
ISSN
Citations 
11619
0302-9743
0
PageRank 
References 
Authors
0.34
0
3
Name
Order
Citations
PageRank
Kiichi Tago122.42
Kosuke Takagi200.68
Qun Jin335146.82