Text emotion classification method based on dual neural network model

Abstract

The invention discloses a text emotion classification method based on a dual neural network model. The method comprises the following steps: step 1, carrying out Chinese word segmentation and stop-word filtering on comment text; step 2, constructing an original training set, and using a provided inversion rule to carry out emotion inversion on the original training set to obtain an inverse training set; step 3, using the original training set and the inverse training set obtained in the step 2 to train the model; step 4, carrying out the same preprocessing as the step 1 on test data to construct anti-sense test text; and step 5, using the emotion classification model to carry out emotion classification on the anti-sense test text. According to the method, the text is represented from a front side of the text, an anti-sense dictionary is utilized at the same time for emotion inversion on the text, and emotion dictionary knowledge can be utilized for more accurate emotion representationon the text; and various text representation neural-network can be used for representing the original text and the inverse text through a dual model framework, the model is very flexible, and accuracyis improved.

Description

technical field [0001] The invention relates to the technical field of natural language processing applications, in particular to a text emotion classification method based on a dual neural network model. Background technique [0002] Under the premise of the rapid development of the Internet, more and more people are included in the Internet. People acquire information, buy goods, and share their lives online, which generates a large amount of text data. Opinion mining and sentiment analysis technology for review texts are not only academic frontier issues and hot research issues in the field of natural language processing and sentiment analysis, but also application fields. An important problem that needs to be solved urgently has immeasurable application value and social significance, but also has great challenges. [0003] In the traditional method, human beings are used to make emotional judgments on texts, which not only requires expensive labor costs, but also has lo...

AI Technical Summary

Problems solved by technology

[0003] In the traditional method, people are used to judge the emotion of the text, which not only requires expensive labor costs, but also has low efficiency.

In recent years, there have been some machine learning methods to try to solve this problem, such as logistic regression model, support vector machine model, neural network model, etc. There is still a lot of room for improvement

Images