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

A Resume Block Classification Method Based on Multi-level Bidirectional Recurrent Neural Network

A two-way loop, neural network technology, applied in branch fields, can solve the problems of low keyword and format feature matching accuracy, large workload, etc., to achieve the effect of improving classification accuracy, improving accuracy, and reducing the scale of training data

Active Publication Date: 2022-07-26
NANJING UNIV OF AERONAUTICS & ASTRONAUTICS
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] In order to solve the shortcomings in the above-mentioned background technology, the present invention proposes a resume block classification method based on a multi-level bidirectional cyclic neural network, in order to fully integrate the format characteristics of resumes, improve the accuracy of resume segmentation and the accuracy of resume block classification rate, so as to solve the problem of low matching accuracy and heavy workload based on keywords and format features

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 Resume Block Classification Method Based on Multi-level Bidirectional Recurrent Neural Network
  • A Resume Block Classification Method Based on Multi-level Bidirectional Recurrent Neural Network
  • A Resume Block Classification Method Based on Multi-level Bidirectional Recurrent Neural Network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0061] In this embodiment, a method for classifying resume blocks based on a multi-level bidirectional cyclic neural network is to first propose a bidirectional long short-term memory cyclic neural network (Bi-LSTM)-based method for the segmentation part of resumes by adopting the idea of ​​sequence labeling. The Resume Segmentation Model (RS) takes each line of text in the resume as the basic granularity, and proposes a format feature code, which is integrated into the feature representation of the line text. All lines of text form a text sequence, and the text The sequence is input into the RS segmentation model to generate a token for each sentence. The tokens are divided into two types: start of block (B) and intra-block (I). For the resume block classification task, a Resume Block Classification Model (RC) based on Bidirectional Gate Unit Recurrent Neural Network (Bi-GRU) is proposed. Each resume block is used as the basic granularity. Sequentially arranged into a sequenc...

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 resume block classification method based on a multi-level bidirectional cyclic neural network, comprising: 1. Resume segmentation: acquiring the training data of the RS model and converting it into vector representation, and then performing the Bi-LSTM layer and the linear layer forward calculation , the model parameters are updated in reverse, and the resume segmentation model is used for prediction, and the segmented resume block sequence is obtained; 2. Resume block classification: The segmented resume block sequence is used as the training data of the RC model, and the Bi‑LSTM and the maximum pooling layer are used. Obtain the CV block feature vector, use Bi-GRU and softmax layer to perform forward calculation to obtain the block category probability distribution, and then use the gradient descent algorithm to update the RC model parameters, so as to use the CV block classification model for prediction. The invention can improve the accuracy rate of resume segmentation and resume block classification, and solve the problems of low accuracy rate and heavy workload of thesaurus establishment based on keyword and format feature matching schemes.

Description

technical field [0001] The invention belongs to the branches of text segmentation, text classification and information extraction in the field of natural language processing in the direction of computer science and technology, in particular to a resume block classification method based on a multi-level bidirectional cyclic neural network. Background technique [0002] Resume information extraction is a technology that uses computer programs to extract semi-structured resume document content into structured content. Through resume information extraction technology, structured resume information can be obtained and stored in a structured storage method, which is convenient for subsequent automatic analysis tools to make further meaningful analysis of these structured resume data. For example, automatic job recommendation, resume screening, resume query, employment recommendation, etc. [0003] The resume information extraction process is generally divided into: resume segment...

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 Patents(China)
IPC IPC(8): G06F40/289G06F16/35G06N3/04G06N3/08
CPCG06F40/289G06F16/35G06N3/08G06N3/047G06N3/048G06N3/045
Inventor 许启强张吉李嘉木
Owner NANJING UNIV OF AERONAUTICS & ASTRONAUTICS