Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Multi-source cross-domain emotion classification method based on MPNet, Bi-LSTM and width learning

A classification method and cross-field technology, applied in neural learning methods, text database clustering/classification, semantic analysis, etc., to achieve a wide range of applications, improve the effect of emotional classification, and strong applicability

Pending Publication Date: 2022-06-21
GUANGDONG UNIVERSITY OF FOREIGN STUDIES
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

), there are still some limitations in practical application

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
  • Multi-source cross-domain emotion classification method based on MPNet, Bi-LSTM and width learning
  • Multi-source cross-domain emotion classification method based on MPNet, Bi-LSTM and width learning
  • Multi-source cross-domain emotion classification method based on MPNet, Bi-LSTM and width learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0059] like figure 1 As shown, the present invention is based on MPNet, Bi-LSTM and a multi-source cross-domain sentiment classification method based on width learning, including the following specific steps:

[0060] Step (1): Manually collect product review data in k source domains and 1 target domain, and then preprocess them (first convert the emojis in the short text into corresponding emotional words, etc.; then, combine A small amount of manual annotation information is used to complete the annotation and storage of the corpus in a machine-based method).

[0061] Step (2): Use the pre-trained model MPNet to pair X sj and X tl The encoding is expressed as follows:

[0062] Using the pre-trained model MPNet for the jth source domain sample X sj and target domain labeled samples X tl The encoding is expressed as follows:

[0063]

[0064]

[0065] where H sj and H tl Respectively for X sj and X tl Using the vector of domain public features generated by MPNe...

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 discloses a multi-source cross-domain emotion classification method based on MPNet, Bi-LSTM and width learning, and the method comprises the following steps: (1) obtaining product comment data of k source domains and one target domain, and carrying out the preprocessing of the product comment data; (2) using a pre-training model MPNet to carry out vectorization representation on a text; (3) carrying out feature extraction on the data by adopting a bidirectional long-short-term memory network Bi-LSTM to obtain field public features; (4) measuring the difference between the distribution of the features Fsj and Ftl by adopting KL divergence; (5) constructing k emotion classifiers based on the DCF; (6) carrying out cooperative training on the k classifiers, and outputting an emotion classification result; according to the method, the advantages of multiple technologies of MPNet, Bi-LSTM and width learning are fused, so that the system can effectively obtain the features of the source field, the emotion classification effect of the target field is effectively improved, and experimental results show that the implementation accuracy of the method provided by the invention is superior to that of a baseline model.

Description

technical field [0001] The invention relates to emotion classification technology, in particular to a multi-source cross-domain emotion classification method based on MPNet, Bi-LSTM and width learning. Background technique [0002] Cross-domain text sentiment classification aims to leverage useful knowledge in the source domain (with sufficient labeled data) to facilitate sentiment classification in the target domain (with little or no labeled data). It not only greatly reduces the dependence on labeled data, but also facilitates sentiment classification for target domains lacking labeled data. Therefore, it has received extensive attention from academia and industry, and has also become a research hotspot in the field of natural language processing. [0003] Not only do product reviews cover many domains, but the size of this data is also extremely unevenly distributed across domains. If sentiment classification of these review texts is required, traditional methods need ...

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
Patent Type & Authority Applications(China)
IPC IPC(8): G06F16/35G06F40/194G06F40/205G06F40/30G06K9/62G06N3/04G06N3/08
CPCG06F16/355G06F40/194G06F40/205G06F40/30G06N3/08G06N3/044G06F18/214G06F18/2411
Inventor 曹丽红彭三城周咏梅
Owner GUANGDONG UNIVERSITY OF FOREIGN STUDIES
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com