Language text processing method and device and storage medium

Abstract

The invention discloses a language text processing method and device and a storage medium, which are used for improving the accuracy of the emotion polarity analysis result expressed by the language text. The language text processing method comprises the following steps: obtaining the language text to be processed; Word segmentation is performed on the language text to be processed to obtain a first word segmentation object, including words obtained by word segmentation and corresponding phonetic alphabet; in accordance with the result of word segmentation, utilizing a vector conversion modelto convert the first word segmentation object into a first word segmentation object vector, the vector conversion model being obtained by training the first word segmentation object contained in the first sample data according to the distance between the first word segmentation object in the first sample data and the affective polarity label of the first word segmentation object; and according tothe first word segmentation object vector, utilizing the emotion polarity prediction model to predict the corresponding emotion polarity type of the language text to be processed, the emotion polarityprediction model being obtained by using the second sample data with emotion polarity labels.

Description

technical field [0001] The invention relates to the technical field of data mining, in particular to a language text processing method, device and storage medium. Background technique [0002] With the rapid development of the Internet, the Internet has become the main means for users to disseminate and obtain various information. In the Internet world, users can use common Internet communication channels, such as forums, microblogs, blogs, post bars, etc., to share information about electronic resources on the Internet, such as games, videos, articles, news, application software, or daily social sentiments, etc. By analyzing the language and text of users’ opinions and attitudes, it is possible to determine the emotional polarity of users’ emotions, opinions, value judgments, and wishes for corresponding electronic resources, which can be used to assist in the analysis of corresponding electronic resources. resources to improve. [0003] In the prior art, the convolutiona...

AI Technical Summary

Problems solved by technology

[0004] Since the existing word vector conversion model mainly considers the distance between words (such as Euclidean distance, etc.) to judge the semantic similarity between them, this leads to the fact that positive and negative emotional words are too close in semantic space, for example, " The words "like" and "hate" in "I like this game" and "I hate this game" have the same context, so the word vectors obtained are very close, but the emotional polarity expressed by these two words is very different. This reduces the accuracy of the resulting sentiment polarity analysis

Images