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

A QA community label recommendation method based on region convolution neural network

A technology of convolutional neural network and Q&A community, applied in the field of tag recommendation in Q&A community

Active Publication Date: 2018-12-25
WUHAN UNIV
View PDF6 Cites 6 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0010] The purpose of the present invention is to provide a label recommendation method based on the regional convolutional neural network for the deficiencies in the label recommendation of the current question-and-answer community, which first performs data preprocessing on the question data set in the question-and-answer community, and then generates sentence vectors by establishing a dictionary , in the convolutional layer, word embedding is performed on each word in each question in the data set, and then the sentence vector is trained with the regional convolutional neural network model, and finally the trained model is used to analyze the Label recommendation for new questions

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
  • A QA community label recommendation method based on region convolution neural network
  • A QA community label recommendation method based on region convolution neural network
  • A QA community label recommendation method based on region convolution neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0062] Embodiments of the present invention will be further described below in conjunction with the accompanying drawings.

[0063] 1) For each sentence o in the domain matrix T i the word W in i , and its corresponding word embedding vector is e(W i ), its corresponding left and right context C l (W i ) and C r (W i ) can be determined by W i-1 The word embedding and left and right context representation, the formula is as follows:

[0064] C l (W i )=f(W l C l (W i ))+W sl e(W i-1 ))

[0065] C r (W i )=f(W r C r (W i+1 ))+W sr e(W i+1 ))

[0066] where W l is the matrix that transforms the current hidden layer to the next hidden layer, W sl is the matrix used to combine the semantics of the current word with the context of the next word. f is a non-linear activation function.

[0067] 2) Word W i representation of x i by word w i and W i The left and right context components are as follows:

[0068] x i =[C l (W i ):e(W i ):C r (W i )] ...

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 relates to a question answering community label recommendation method based on a region convolution neural network, specifically, the method involves the data preprocessing of the question data set in the question answering community, then, by building dictionary to generate sentence vectors, every word in the data set is embedded in the convolution layer, and then the sentence vectors are trained with the regional convolution neural network model; finally, the trained model is used to recommend the new questions in the question-answer community. The method has the following outstanding characteristics and advantages: firstly, the bidirectional circulating convolution layer is introduced to express the words in combination with the context of the words, which can better reflect the relationship between the words in the sentence, and the bidirectional circulating structure can accurately grasp the context of the words; secondly, the ability of convolution neural network todeal with large images makes it possible to deal with large data sets; thirdly, convolution neural network has good migration in image processing, which makes it have good migration in label recommendation.

Description

technical field [0001] The invention relates to a tag recommendation method for a question answering community based on a regional convolutional neural network. Background technique [0002] With the rapid development of the Internet, more and more IT enthusiasts seek help, share experience, and learn new technology knowledge on the Internet, thus various developer question-and-answer communities, software information sites such as StackOverflow and Freeecode have emerged It can provide information sharing and communication for developers all over the world. [0003] To facilitate proper categorization and efficient searching, developers need to provide tags for their publications. However, tagging is an inherently uncoordinated process that depends not only on a developer's understanding of their own posts, but also on other factors, including the developer's English language skills and knowledge of existing posts. [0004] As a result, developers keep creating new tags e...

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): G06F17/30G06N3/04
CPCG06N3/045
Inventor 刘进周平义储玮李兵崔晓晖陈旭施泽洋彭新宇赵发凯
Owner WUHAN 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