Text sentiment analysis method

A sentiment analysis and text technology, applied in the field of text analysis, which can solve the problems such as the distance between sentiment words and target words, the lack of emotional tendencies, and the difficulty in expressing the meaning of sentences accurately and efficiently.
CN113761941AActive Publication Date: 2021-12-07SOUTH CHINA NORMAL UNIVERSITY
7 Cites 3 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
SOUTH CHINA NORMAL UNIVERSITY
Publication Date
2021-12-07

Smart Images

  • Figure 1
    Figure 1
  • Figure 2
    Figure 2
  • Figure 3
    Figure 3
Patent Text Reader

Abstract

The invention provides a text sentiment analysis method, which comprises the following steps: inputting a general semantic matrix, a structural semantic matrix and a sentence initialization representation into a graph convolutional neural network, a cross network and an activation function to obtain a general semantic syntactic representation and a structural semantic syntactic representation of a sentence; calculating a target structure semantic syntactic representation and a target general semantic syntactic representation in combination with the general semantic hidden representation and the structural semantic hidden representation of the sentence, splicing the target general semantic syntactic representation and the target structure semantic syntactic representation, and shielding non-attribute words; and obtaining attribute word feature representation of each word in a sentence, then calculating attention weight of each word in the sentence according to the word embedding representation and each attribute word feature representation, and outputting sentence sentiment polarity based on the word embedding representation and the attention weight of each word. According to the text sentiment analysis method, the sentiment polarity of the sentences in the text can be efficiently and accurately analyzed.
Need to check novelty before this filing date? Find Prior Art

Description

technical field

[0001] The invention relates to the technical field of text analysis, in particular to a text sentiment analysis method. Background technique

[0002] Fine-grained sentiment analysis (ABSA) is an important task in natural language processing. For a given (sentence, attribute word) pair, it aims to judge the attribute word according to the context information related to the specific attribute word in the sentence. emotional polarity (eg, positive, neutral, negative). With the continuous development of the Internet and big data, sentiment analysis technology that can automatically extract useful information from a large number of texts has not only attracted the attention of many academic circles, but has also been widely used in daily business activities, especially in social media. , e-commerce and social public opinion analysis and other fields.

[0003] In the existing fine-grained sentiment analysis process, the main goal is to accurately find sentiment ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More