Text sentiment analysis method and system based on graph convolution network and electronic device

Abstract

The invention relates to a text sentiment analysis method and system based on a graph convolution network and an electronic device. The method comprises the steps: word segmentation being conducted onan input text sequence; converting each segmented word into a corresponding word embedding according to the sequence of the text sequence; extracting a forward semantic feature and a reverse semanticfeature of each word embedding, and combining the forward semantic features and the reverse semantic features at the same positions to obtain a context semantic feature of each word embedding; according to the context semantic feature of each word embedding, calculating a semantic relationship value between any two word embedding to obtain a connection matrix; analyzing a dependency syntax tree of the text sequence according to the connection matrix; performing graph convolution operation by taking the dependency syntax tree as a graph to obtain dependency vectors of ROOT nodes of the dependency syntax tree; and performing sentiment polarity classification scoring on the dependency vector of the ROOT node position in the dependency syntax tree, and determining the sentiment polarity category of the text sequence.

Description

technical field [0001] The present application relates to the technical field of text sentiment analysis, in particular, to a text sentiment analysis method, system and electronic device based on a graph convolutional network. Background technique [0002] Sentiment analysis technology gradually emerged with the rapid development of the Internet in the early 20th century, and has gradually expanded from the field of academic research to the field of industrial applications. As a text classification task, text sentiment analysis used a lexicon-based approach in the early stage, pre-constructing a large enough sentiment lexicon, and then using rules to determine the emotional tendency of the text. However, the construction process of the emotional dictionary requires manual sorting of various types of words, and the continuous emergence of new words requires continuous maintenance of the emotional dictionary, resulting in a huge human investment for this type of method; at the...

AI Technical Summary

Problems solved by technology

[0003] The purpose of this application is to solve the problem that the existing text sentiment analysis technology has a weak ability to extract long-distance semantic relations, cannot directly model the semantic relations between non-adjacent words, and has poor flexibility in domain transfer The problem

Images