Design And Implementation Of IT Resume Scoring Model

With the rapid development of information technology in recent years,more and more IT positions emerge.A large number of non-technical enterprises also put into the wave of building technical teams for business development.Recruiters don’t usually have a lot of these skills,and it’s a daunting task to sift through a huge number of resumes from different recruitment platforms every day that might fit the bill.Therefore,it is necessary to develop a resume screening system specifically for IT recruitment to reduce the pressure of the personnel department,so that recruiters can efficiently screen out qualified resumes.Compared with traditional positions,the biggest difficulty in resume screening for IT positions is to judge the matching degree of technology-related skills.This thesis focuses on the core dimension of technology stack,supplemented by the traditional evaluation dimension,and carries out the following aspects of work:First,use data mining technology to build a basic database that can support the resume scoring model.On the premise of following the crawler agreement of recruitment platform,crawl enough job data and resume data.The job data is used to construct the technology stack dictionary and generate the scoring model by using the collaborative filtering algorithm,while the resume data is used to test and optimize the scoring model.Secondly,the resume scoring model is constructed by combining NLP technologies such as text segmentation algorithm,similarity calculation method and collaborative filtering algorithm.This model supports online construction and updating and can solve the screening problem of various new technical job resumes.Through the operation learning of the corpus of more than ten thousand technical positions,the basic database of several conventional positions has been formed,and the test has been carried out in the resume database of tens of thousands of jobs mixed with technical and non-technical categories,showing good generalization ability.Third,turn the scoring model into a web site system that can be used in a production environment.Use PHP,MySQL,Vue and other development technology to turn the relevant algorithm into an operational man-machine interface,hide the underlying logic details,only by a few simple steps to complete the resume screening work.The best practice for matching people and posts is to reduce ineffective actions on both sides,allowing companies to find the right talent and preventing poorly matched candidates from wasting time in unrelated posts.The development of resume screening system has realized the achievement transformation from algorithm to platform,which really lightens the burden of recruiters,improves recruitment efficiency and saves recruitment costs.