Text emotion recognition method and device, electronic equipment and storage medium

Abstract

The invention relates to a text emotion recognition method and device, electronic equipment and a storage medium. The method comprises the steps that a character matrix of a to-be-recognized text is determined, wherein the character matrix is composed of character vectors corresponding to characters in the to-be-recognized text; a position vector of the to-be-recognized text is determined, the position vector being used for representing a position of a target character in the to-be-recognized text, and the target character being a character matching a target emotion in the characters; a text matrix formed by the character matrix and the position vector is input into a pre-trained emotion recognition model, the emotion recognition model is used for extracting context information of the to-be-recognized text from the text matrix, and the matching degree of the to-be-recognized text and the target emotion is determined and output according to the information vector corresponding to the context information. According to the scheme, the context information of the to-be-recognized text can be extracted, and the real semantics of the text can be fully understood, so that the position vector can ensure the accuracy of emotion recognition.

Description

technical field [0001] The present disclosure relates to the field of machine learning, and in particular to a text emotion recognition method, device, electronic device and storage medium. Background technique [0002] Texts such as articles, sentences, and comments can usually reflect the author's emotional tendency, such as the likes and dislikes of a certain thing, the preference for a certain product, etc. Identifying the emotional tendency of the text is the current stage of Natural Language Processing (NLP). ) technology is an important means. At this stage, the industry uses a variety of methods to realize the emotion recognition of text. [0003] One type of scheme is to use sentiment dictionary matching or a classification model based on machine learning to perform sentiment recognition on text. Such schemes usually traverse the text to be recognized from front to back according to the pre-established sentiment lexicon, and then determine the emotional tendency o...

AI Technical Summary

Problems solved by technology

However, this kind of scheme not only depends on the construction method of the emotional dictionary and the richness of the vocabulary to a large extent, but also can only analyze the emotional tendency of the text from the word granularity and cannot understand the semantics of the text, so the accuracy is low and the applicable scenarios are limited.

[0004] Another type of scheme uses SVM (Support Vector Machine, Support Vector Machine) or naive Bayesian and other machine learning-based classification models to carry out emotion recognition on text, but this type of model has a strong dependence on the selection of word features, and because of the use of Therefore, word vectors are also difficult to fully express semantic information, so it is difficult to use and the accuracy needs to be improved

Images