Research And Implementation Of Person-job Matching Algorithm Based On Deep Learning

Online recruitment services are rapidly changing the traditional recruitment pattern in the job market.There are hundreds of millions of registered users with resumes,and tens of millions of recruitment information real and available on the net.With the continuous growth of online recruitment data,the selection of recruiters and job seekers has also increased the difficulty,so automatically matching suitable job and resumes has become an important task.Traditional person-job matching requires a lot of preliminary preparations.With the development of machine learning,researchers have applied deep learning and neural networks to resume-job matching tasks and conducted a lot of research on topics.This paper mainly studies the representation learning and collaborative learning methods in the process of job-resume intelligent matching,and has achieved the following research results.(1)Proposing a Person-Job Fit network Based on Bidirectional Preference Information Existing researches on person-job matching mainly focus on learning good representations of job descriptions and resume texts with comprehensive matching structure.This article assumes that learning the preference of recruiters and job seekers from previous interview histories will bring benefits,and hopes that this preference will help improve the degree of match between the job post and resume.To this end,this paper proposes a novel matching network based on preference modeled.The key idea is to explore the potential preference of both sides given the history of all interviewed candidates for a certain job posting and all the job application records for a particular talent.More specifically,this article proposes a historical memory module to learn the potential preference representation by interacting with both work and resume.Then the preference is integrated into the matching framework as an end-to-end learnable neural network.Based on the real data of the online recruitment platform,the experimental results show that compared with a series of state-of-the-art methods,the proposed model can improve the performance of job-resume matching.In this way,it is demonstrated that recruiters and job seekers indeed have preference,and this preference can improve job-resume matching in the job market.(2)Proposing a Job-Resume Matching network Based on Multi-Perspective Collaborative learning In previous studies,the matching task is usually transformed into a supervised text matching problem.When the labeled data is sufficient,supervised learning is powerful.However,on online recruitment platforms,the job-resume interaction data is sparse and noisy,which affects the performance of job-resume matching algorithm.In order to solve these problems,this paper proposes a multi-view collaborative learning network based on sparse interactive data for job-resume matching.The matching network is mainly composed of two parts,namely the text-based matching model and the relation-based matching model.The two parts capture semantic compatibility in two different views and complement each other.In order to alleviate the challenge from sparse and noisy data,this paper designs a joint teaching mechanism to combine the two components.First,the two components share the learned parameters or representations so that the original representations of each component can be enhanced.More importantly,this article allows the two components to reduce the impact of noise in training data by selecting more reliable training samples.The two steps respectively focus on data representation enhancement and data quality enhancement.Compared with the pure text-based matching model,this method can learn better data representations from limited or even sparse interactive data,which is stronger resistance to noise in the training data.We have conducted experiments with the recruitment data of different professional fields.Then,the results show that the proposed model can outperform cutting-edge jobresume matching methods.Finally,the improved job-resume matching algorithm of this article is applied to the online recruitment system.In addition,the intelligent job-resume matching module of this system is designed and implemented.