Method and system for text sentiment analysis and processing

Abstract

The invention relates to a method and system for text sentiment analysis and processing. The method comprises the steps that word segmentation is conducted on a text; word vector training is conducted on segmented words of the text, so a binary file is obtained; sentiment characteristic word groups are extracted from the binary file, and syntax characteristic information and sentiment characteristic information are acquired from the word groups; characteristic integration is conducted on the syntax characteristic information and the sentiment characteristic information, so text characteristics containing syntaxes and sentiment information are obtained; the word vectors and the sentiment characteristic information in the binary file are integrated, so word vectors containing the sentiment information are obtained; the word vectors are extracted, so semantic characteristics containing the sentiment information are obtained; and the text characteristics containing the syntax and sentiment information are integrated with the semantic characteristics containing the sentiment information, so grammar information, semantic information, syntax information and sentiment information of the text can be obtained. According to the invention, the problem in the prior art, that the extracted characteristics cannot contain the semantic information, syntax information and sentiment information at the same time, can be solved; and results obtained are highly accurate.

Description

technical field [0001] The invention relates to the technical field of data analysis, in particular to a method and system for text sentiment analysis and processing. Background technique [0002] With the vigorous development of communication Internet technology and the advent of the web3.0 era, the Internet has gradually shifted from a one-way communication model to an open architecture concept that takes users as the center and enables everyone to participate. Internet users no longer just read web pages but can write their own opinions on web pages, changing from passively accepting Internet information to actively accepting and disseminating Internet information. Internet applications such as microblogs, blogs, forums, news and product reviews have produced a large number of user-participated, subjective and valuable commentary information on people, events, and products. By analyzing the emotional direction of these comments, the government can better understand the p...

AI Technical Summary

Problems solved by technology

[0008] In the existing technology, when analyzing the emotional polarity of text, the extracted features cannot contain semantic information, syntactic information and emotional information at the same time, so the extracted results are not accurate enough

Images