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

An emotion polarity detection method based on multi-domain data set joint embedding

A technology of polarity detection and data collection, which is applied in the field of service computing to solve the effect of performance degradation

Pending Publication Date: 2019-06-14
SHANDONG UNIV OF SCI & TECH
View PDF3 Cites 7 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] In view of the limitations of the existing emotional polarity detection methods, the purpose of the present invention is to provide a sexy polarity detection method based on joint embedding of multi-domain data sets to solve the problems in the above-mentioned background technology

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
  • An emotion polarity detection method based on multi-domain data set joint embedding
  • An emotion polarity detection method based on multi-domain data set joint embedding
  • An emotion polarity detection method based on multi-domain data set joint embedding

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0060] Below is the specific embodiment of the application of the present invention:

[0061] Collect dataset documents, dataset documents include three parts: Amazon domain adaptation dataset and SemEval2013 and 2016 datasets. The Amazon domain adaptation dataset contains four domains, each of which contains 1600 training and 400 testing. The SemEval 2013 and 2016 data sets contain 3547 training data and 1262 test data, 4124 training data and 8005 test data.

[0062] Perform step 1 to label positive data labels as "1" and negative data labels as "0" in the six domains. Use the one-hot tool to convert the label into a one-hot vector;

[0063] Execute step 2, use the Word2Vec model to generate word embeddings for multiple source domain and target domain pairs. The mapping between the source domain and the target domain is established through a dictionary. Then two linear projection matrices are used to create a mapping from the original vector space to the dual-domain space...

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 an emotion polarity detection method based on multi-domain data set joint embedding, which belongs to the technical field of service computing, and utilizes the advantages of across-domain data set to carry out emotion polarity detection on data sets in different fields. Inspired by the latest progress of emotional analysis performed by the cross-domain data set, the invention provides a new visual angle, and the domain adaptation problem of the data set is taken as an embedded projection task. The model of the invention takes two single-field embedding spaces as inputand projects the two single-field embedding spaces to a double-field space through learning, and the space is subjected to joint optimization to predict emotion polarity. An experiment is carried outon a plurality of source domain and target domain pairs by using the Amazon domain adaptive data set and the SemEval 2013 and 2016 data sets so as to carry out sentiment classification. The result analysis shows that the model provided by the invention is equivalent to the most advanced method in the similar fields, and is better in performance in the fields with different heights.

Description

technical field [0001] The invention relates to the technical field of service computing, in particular to a sexy polarity detection method based on joint embedding of multi-domain data sets. Background technique [0002] One of the main limitations of current sentiment polarity detection methods is that they are sensitive to domain differences. This causes the classifier to perform poorly in new domains after training. Domain adaptation techniques provide a solution to reduce variance and make models perform well across multiple domains. The two main approaches to domain adaptation for sentiment analysis are pivot-based methods, which augment the feature space with domain-independent features learned unsupervised, and data and autoencoder methods, which attempt to create an A good general mapping to latent hidden spaces. While pivot-based domain adaptation methods are powerful, they generally outperform autoencoder methods. However, both domain adaptation methods lead t...

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): G06F17/27G06K9/62G06N3/08
Inventor 田刚王琦博刘鹏飞孙承爱
Owner SHANDONG UNIV OF SCI & TECH
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