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

Emotion classification method and system

A sentiment classification and sentiment technology, applied in the computer field, can solve the problems of not being able to capture full text information, affecting the accuracy rate, and high computational complexity

Inactive Publication Date: 2016-09-07
SHENZHEN UNIV
View PDF4 Cites 19 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to provide a method and system for sentiment classification, which aims to solve the problem of the inability to capture the full text due to the high computational complexity of word-level training in the prior art, the lack of labels of internal nodes affecting the accuracy, and the one-way propagation. information problem

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 classification method and system
  • Emotion classification method and system
  • Emotion classification method and system

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0025] figure 1 The implementation flow of the emotion classification method provided by the first embodiment of the present invention is shown. For the convenience of description, only the parts related to the embodiment of the present invention are shown, and the details are as follows:

[0026] In step S101 , in the phrase binary tree, recursively go up layer by layer from the leaf node, and calculate the vector of each node, and the vector of the node is a vector based on the phrase level.

[0027] In the embodiment of the present invention, the leaf nodes of the phrase binary tree are phrases rather than words, therefore, phrase-level vectors need to be obtained as initial input data. First, word-level vectors need to be obtained, and then the word vectors are calculated in a certain combination to obtain phrase vectors, which are the vectors of the nodes of the phrase binary tree. In practical applications, the semantic word embedding representation can be learned throu...

Embodiment 2

[0052] image 3 The implementation process of the emotion classification method provided by the second embodiment of the present invention is shown. For the convenience of description, only the parts related to the embodiment of the present invention are shown, and the details are as follows:

[0053] In step S301, the phrase dependency tree is converted into a phrase binary tree.

[0054] Figure 4 It shows the implementation process of converting the phrase binary tree in the sentiment classification method provided by the second embodiment of the present invention, converting the phrase dependency tree into a phrase binary tree, including:

[0055] In step S401, the phrase dependency tree is parsed layer by layer from the bottom to obtain the triplet structure in each layer.

[0056] In this embodiment, in the process of constructing the phrase binary tree, each embedded structure T in the phrase dependency tree is stored i , according to the phrase dependency tree struc...

Embodiment 3

[0069] Figure 5 A schematic structural diagram of the emotion classification system provided by Embodiment 3 of the present invention is shown. For the convenience of illustration, only the parts related to the embodiment of the present invention are shown, including: vector calculation unit 51, emotion label determination unit 52, feedback vector Calculation unit 53 and sentiment classification unit 54, wherein:

[0070] The vector calculation unit 51 is configured to recurse layer by layer from the leaf node in the phrase binary tree to calculate the vector of each node, and the vector of the node is a vector based on the phrase level.

[0071] The emotional label determination unit 52 is used to determine the emotional label of the node by calculating the similarity with a reference word, the reference word is an emotional word with strong positive and negative, and the emotional label is an emotional tendency value.

[0072] The feedback vector calculation unit 53 is con...

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 is suitable for the technical field of computers, provides an emotion classification method and system. The emotion classification method includes: performing recursion upward from a leaf node layer-by-layer in a phrase binary tree, and calculating a vector of each node, wherein the vector of each node is a phrase level-based vector; determining an emotion label of each node by calculating the similarity between a reference word and the vector, wherein the reference word is an emotion word with strong positive / negative, and the emotion label is an emotion tendency value; performing recursion downward from a root node layer-by-layer, and calculating a feedback vector of each node, wherein the feedback vector of each node is phrase level-based vector; and performing emotion classification through a classifier function according to the vector, the feedback vector and the emotion label of each node. The phrase binary tree is constructed, phrase level operation is performed, the emotion label of each node is acquired, and then classification is performed according to the vector, the feedback vector and the emotion label of each vector, full text information can be captured through two-way communication, and the accuracy of classification can be improved.

Description

technical field [0001] The invention belongs to the technical field of computers, and in particular relates to an emotion classification method and system. Background technique [0002] Text sentiment classification is to analyze and process subjective texts with emotional color, summarize and infer the emotional tendency of the text, which can be divided into chapter level, paragraph level, sentence level and word level according to the different granularity. At present, there are many researches on sentiment analysis of English text. For example, Turney uses unsupervised learning method to detect the polarity of product review data; Pang uses machine learning method to classify movie reviews. Both of these works are on document-level text The data are classified into two categories. [0003] Traditional sentiment analysis methods are mainly based on word-level feature expressions, and word embedding representations are used for sentence-level or document-level sentiment a...

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): G06F17/30G06F17/27
CPCG06F16/35G06F40/289
Inventor 傅向华徐莹莹
Owner SHENZHEN UNIV
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