Open AccessMultifaceted sentiment analysis of public comments on the Dianping.com
Author(s) -
Liying Sui,
Luning Shang,
Xiaomao Guo,
Dexue Zhang
Publication year - 2020
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1550/3/032052
Subject(s) - computer science , artificial intelligence , sentiment analysis , relevance (law) , semantics (computer science) , classifier (uml) , mechanism (biology) , artificial neural network , set (abstract data type) , natural language processing , term (time) , machine learning , epistemology , philosophy , physics , quantum mechanics , political science , law , programming language
Emotional analysis of text has always been a hot topic in natural language research. In view of the long-term dependence of recurrent neural networks and the fact that most models do not consider the correlation between input and output, this paper proposes a bidirectional LSTM model based on attention mechanism to judge the comment emotion. This method vectorizes the semantics of comments into the LSTM model, improves the relevance of input and output through the attention mechanism, fuses the aspect category and aspect Term, and outputs the results through the classifier. The experimental results of Dianping.com review data set provided by AI Challenger competition show that the improved method adopted in this paper is better than the common deep learning method.