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

Emotion cause pair extraction method based on emotion embedding

An emotion and reason technology, applied in the field of emotion reason extraction based on emotion embedding, can solve problems such as failure to use

Active Publication Date: 2022-07-29
TIANJIN UNIV
View PDF6 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

These studies completely ignore the fact that there is a causal relationship between emotions and causes, failing to take advantage of their mutually indicative properties

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
  • Emotion cause pair extraction method based on emotion embedding
  • Emotion cause pair extraction method based on emotion embedding
  • Emotion cause pair extraction method based on emotion embedding

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0046] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention, but not to limit the present invention.

[0047] like figure 1 As shown, the method for extracting emotional cause pairs based on emotional embedding of the present invention utilizes the model for emotional cause pair extraction based on emotional embedding to achieve emotional cause pair extraction, and the implementation steps of the model are as follows:

[0048] Step 1: Word Embedding Coding

[0049] The input of the model obtains the time series representation of the text through word embedding s={s 1 ,s 2 ,…,s n },

[0050] S2. Emotional coding

[0051] The sentiment word-level feature representation is obtained through the first-layer word-level Bi-LSTM network

[0052]

[0053] in, is a word-level Bi...

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 reason pair extraction method based on emotion embedding, which is characterized in that emotion reason pair extraction is realized by utilizing an emotion reason pair extraction model based on emotion embedding, and the model extraction is realized through word embedding, emotion coding, reason coding, emotion reason pair coding and emotion reason pair prediction. Word embedding representation of the input text is obtained through word embedding; emotion clause-level feature representation is obtained through emotion coding; obtaining cause clause-level feature representation through cause coding; then utilizing the emotion reason pair codes to obtain emotion reason pair feature codes; and finally, utilizing the probability distribution on a softmax layer output label to realize emotion reason pair extraction. According to the method, the emotion reason pair extraction capability can be improved by utilizing the causal relationship and the mutual indication characteristic between the emotion and the reason.

Description

technical field [0001] The invention relates to the technical field of emotional cause analysis, in particular to a method for extracting emotional cause pairs based on emotional embedding. Background technique [0002] The Internet has become more and more popular, people's lives are more and more closely connected with the Internet, and more and more people have begun to obtain various information and express their views and opinions through the Internet. Text data is one of the highest quality and largest quantity of new data that is continuously generated on the Internet every day. On the other hand, text data is also a data type that is most in line with people's language habits, because it is one of the most important carriers for people to obtain information and express emotions. How to analyze these massive text data to obtain useful information has very important research value. The sentiment analysis of text data is a direction of specific analysis. However, the ...

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): G06F16/36G06F40/211G06F40/284G06N3/04G06N3/08
CPCG06F16/367G06F40/211G06F40/284G06N3/08G06N3/044Y02D10/00
Inventor 饶国政李志威
Owner TIANJIN UNIV