Target-specific Sentiment Polarity Classification Method

Abstract

The present invention provides a method for classifying emotional polarity of a specific target, comprising: obtaining a hidden state vector corresponding to a context, performing multi-head self-attention coding on it, and obtaining contextual semantic information coding; a graph convolutional neural network based on a combined gating mechanism , to extract the syntactic vector corresponding to the context; through the weight vector corresponding to the specific target, the contextual syntactic information code related to the specific target is obtained; the contextual semantic information code and the contextual syntactic information code after splicing are encoded by multi-head self-attention to obtain the contextual semantic syntax Information encoding; the context semantic information encoding and the context semantic syntax information encoding are averaged and pooled and then spliced ​​to obtain the feature representation of a specific target; the feature representation of a specific target is input into the preset emotional polarity classification function to obtain the emotional polarity of a specific target sex classification probabilities. Compared with the prior art, the present invention can effectively filter invalid syntactic information and improve the accuracy of sentiment classification of a specific target.

Description

technical field [0001] The invention relates to the technical field of natural language processing, in particular to an emotion polarity classification method for a specific target. Background technique [0002] Sentiment analysis is an important task in Natural Language Processing (NLP), and its purpose is to analyze subjective text with emotional color. Among them, target-specific sentiment analysis belongs to fine-grained sentiment analysis, which is different from traditional sentiment analysis, and its purpose is mainly to identify the emotional polarity of a specific target in a sentence. [0003] At present, there are many methods combining neural network and attention mechanism to solve the problem of emotional analysis of specific targets. Although these methods can overcome the defects of shallow learning models and distinguish the importance of different words for specific target emotional analysis tasks, However, there are still the following problems: on the on...

AI Technical Summary

Problems solved by technology

[0003] At present, there are many methods combining neural network and attention mechanism to solve the problem of emotional analysis of specific targets. Although these methods can overcome the defects of shallow learning models and distinguish the importance of different words for specific target emotional analysis tasks, However, there are still the following problems: on the one hand, the semantic information of the context cannot be fully captured, and attention is paid to the long-distance information dependence and information parallel computing; accuracy

Images