Text sentiment classification method

Abstract

The invention is applicable to the technical field of sentiment classification, and provides a text sentiment classification method, which comprises the following steps: performing dynamic word vector modeling on a target text based on a Bert model, and inputting dynamic word vector data into a CNN channel of a preset dual-channel neural network model for feature learning to obtain a first feature vector; obtaining topic information of the target text, performing static word vector modeling on the topic information based on a word2vec model, and inputting static word vector data into a GRU channel of a preset dual-channel neural network model for feature learning to obtain a second feature vector; splicing the first feature vector and the second feature vector to obtain a third feature vector; and processing the third feature vector through a self-attention mechanism, and performing sentiment classification on the processed third feature vector based on a preset classifier model. The sentiment classification method and device can improve the sentiment classification accuracy of the target text.

Description

technical field [0001] The invention belongs to the technical field of sentiment classification, and in particular relates to a text sentiment classification method. Background technique [0002] Sentiment classification is also called opinion mining, tendency analysis, etc. It refers to the use of text mining, natural language processing and other technologies to identify and extract subjective information in comment texts, and to obtain the opinions and attitudes of the analysis object on a certain topic or a certain text. [0003] At present, using deep learning methods and attention mechanism technology to classify the sentiment of review texts has become a new research hotspot. However, the inventors of the present application found that after the existing emotion classification method has trained a language model, the word vector of each word is fixed. When the word vector is used later, no matter what the input sentence is, the word vector Neither has changed and can...

AI Technical Summary

Problems solved by technology

[0004] In view of this, the embodiment of the present invention provides a text sentiment classification method to solve the problem that the sentiment classification method in the prior art cannot deal with polysemy and poor classification accuracy

Images