View-level text sentiment classification system and method based on a graph convolutional neural network

Abstract

The invention relates to a view-level text sentiment classification system and method based on a graph convolutional neural network, and the system comprises a text preprocessing module which is used for carrying out the feature processing of a view-level text; a text semantic information acquisition module used for capturing a bidirectional semantic dependency relationship of the text; an attention coding module used for capturing global internal correlation of a text word sequence and carrying out further information integration; a graph convolutional neural network module which directly acts a GCN on a sentence dependency relationship tree to model a sentence structure, and can spread context and dependency information from viewpoint words to viewpoint words; and a sentiment category output module used for obtaining a final sentiment classification result of the text by using the classification function. The effect of the graph convolutional neural network can be effectively exerted, the sentence structure can be modeled through the semantic dependency tree by utilizing the graph convolutional neural network, and better text emotion feature representation is obtained.

Description

technical field [0001] The present invention relates to the field of text analysis, in particular to a method and system for text sentiment classification at the perspective level based on a graph convolutional neural network. Background technique [0002] In the past, the object of sentiment analysis was mainly sentences or documents, and achieved good results. However, in real application scenarios, the simple use of the sentiment (positive, negative and neutral) of a piece of text cannot accurately express the true sentiment of the text. Take the comment "This mobile phone is very cheap, but the pixels are not high" as an example. This comment is an overall description of a mobile phone, but sentiment analysis at the document level or sentence level cannot determine the emotional polarity of the comment. View-level text sentiment analysis is a fine-grained task in sentiment analysis, which aims to study the emotional polarity expressed by each view word in the text. The ...

AI Technical Summary

Problems solved by technology

[0002] In the past, sentiment analysis mainly focused on sentences or documents, and achieved good results. However, in real application scenarios, the sentiment of a piece of text is simple to use (positive , negative and neutral) cannot accurately express the true emotion of the text

Take the comment "This mobile phone is very cheap, but the pixels are not high" as an example. This comment is an overall description of a mobile phone, but sentiment analysis at the document level or sentence level cannot determine the emotional polarity of the comment.

Images