Design And Implementation Of A Job Applicant Matching System Based On Deep Learning

In recent years,with the rapid development of the Internet,human resources management has become more and more dependent on Internet.More companies choose online recruitment and online seminars to carry out corporate recruitment and promotion work,and more job seekers also start to choose to submit their resumes as well as interviews through Internet channels.However,there are still some shortcomings in the current enterprise recruitment.First of all,companies and job seekers asymmetric access to information,the job market between companies and job seekers information is not transparent.Job seekers can not quantify the specific level of the company,the company also does not have an intuitive positioning of all aspects of the conditions of job seekers.Job seekers can not clearly understand their own strengths and weaknesses.Secondly,the focus of existing job boards is to provide job seekers and companies with services related to resume delivery,lack of macro-level to provide job seekers and companies with a visual analysis of the general environment of the recruitment market.Therefore,both job seekers and companies lack data-level perception of the market environment.Finally,the existing recruitment websites are more concerned with the process of job seekers actively searching for jobs for delivery,and less concerned with the active and personalized job recommendation function for job seekers.Based on the above analysis,this paper implements a deep learning-based job applicant matching system.The system uses Echarts for data visualisation and Flask framework for background construction,and data is stored through csv files,My SQL and configuration files.Based on preliminary background research and demand analysis,the system is divided into four modules: crawler management,recruitment market environment analysis,user management and job search management according to business logic.Each module is interoperable to provide rich job search and recruitment services for job seekers and corporate users.The system updates the background data through regular crawlers to display the realtime analysis results of the recruitment market environment to users in a visualized form,and digs out the relationship between different factors such as job,salary,region and experience and education.Moreover,the system quantifies the level of job seekers and job competency requirements in various dimensions in the form of radar diagrams,and designs and produces job seeker portraits and job competency requirements mapping.Finally,the system is based on deep learning algorithms to match job seekers and make intelligent recommendations for job seekers to help them apply for jobs more accurately.The job intelligence recommendation algorithm part uses BiLSTM and Attention mechanism to construct a word-level person-job matching neural network APJFNN.The model mainly utilizes the information from rich historical job application data and consists of three parts: Word-level Representation,Hierarchical Ability-aware Representation and Person-Job Fit Prediction.Based on the word-level semantic representations of job requirements and job applicants' resumes,four hierarchical abilityaware Attention mechanisms are designed to represent their different levels of ability representations.Finally,the learned job requirements and resume representations are fed back to the person-job matching prediction to evaluate the matching degree between them.I have independently completed the requirement analysis,design,implementation and testing of the system.All the module functions of this project have been basically implemented and the system is running normally.