Unlock instant, AI-driven research and patent intelligence for your innovation.

Specific target sentiment analysis method for fusion word shielding data enhancement and adversarial learning

A specific target, data shielding technology, applied in the direction of electrical digital data processing, text database query, natural language data processing, etc., to achieve the effect of improving performance

Active Publication Date: 2021-11-30
CHONGQING UNIV OF TECH
View PDF8 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Due to the lack of publicly available ABSA data sets, it is difficult to make the model have sufficient generalization ability, model robustness and model efficiency

Method used

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
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Specific target sentiment analysis method for fusion word shielding data enhancement and adversarial learning
  • Specific target sentiment analysis method for fusion word shielding data enhancement and adversarial learning
  • Specific target sentiment analysis method for fusion word shielding data enhancement and adversarial learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0077] Embodiments of the present invention are described in detail below, examples of which are shown in the drawings, wherein the same or similar reference numerals designate the same or similar elements or elements having the same or similar functions throughout. The embodiments described below by referring to the figures are exemplary only for explaining the present invention and should not be construed as limiting the present invention.

[0078] 1. Related work

[0079] ABSA, also known as fine-grained sentiment analysis, has always been a research hotspot at home and abroad. The main task of this task is to determine the sentiment classification of a specific target entity in each sentence. Although the current research has achieved good results using deep learning methods, there are few training data that lead to weak generalization ability, and the accuracy and robustness of the model. Sexual incompatibility and other issues.

[0080] Traditional deep learning method...

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

PUM

No PUM Login to View More

Abstract

The invention provides a specific target sentiment analysis method for fusion word shielding data enhancement and adversarial learning, which comprises the following steps of: S1, performing synonym replacement and random word insertion on sentences in a target entity shielding manner to generate effective samples, and fusing the effective samples with original samples to realize word shielding data enhancement; s2, constructing an adversarial learning specific target sentiment classification model based on BERT-BASE, and training the sentiment classification model by using a clean sample and an confrontation sample together, so that the model has an adversarial defense function; and S3, performing adversarial learning on the original sample and the sample after data enhancement. According to the method, data enhancement and adversarial training are adopted, high robustness is achieved, and a better result can be obtained.

Description

technical field [0001] The invention relates to the field of natural language processing, in particular to a specific target sentiment analysis method which combines word masking data enhancement and confrontation learning. Background technique [0002] With the rapid development of social media (Microblog, Twitter, Facebook, etc.), sentiment analysis has become an extremely important task. Aspect-based Sentiment Analysis (ABSA) is a basic task in the field of text classification. The purpose is to use deep learning and natural language processing (Natural Language Processing, NLP) technology to analyze the fine-grained emotional tendency of online social network text data, which is convenient for users. Have a clear understanding of the hidden sentiment polarity (Sentiment Polarity) and attitude (Attitude) of specific entities (Aspect) in online social network comment data. A sentence contains one or more entities, and each entity has a different sentiment polarity. For ex...

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

Application Information

Patent Timeline
no application Login to View More
IPC IPC(8): G06F40/211G06F40/247G06F40/284G06K9/62G06N3/02G06F16/33
CPCG06F40/247G06F40/211G06F40/284G06F16/3344G06N3/02G06F18/214Y02D10/00
Inventor 刘小洋代尚宏高绿苑
Owner CHONGQING UNIV OF TECH