Research On Bidirectional Matching Recommendation Algorithm Of Resume And Recruitment Information Based On Text Similarity

In the era of rapid development of Internet information,the traditional offline recruitment mode is slowly fading out of people’s vision.Today’s recruitment is mainly promoted through the network.The network recruitment mode has the advantages of wide coverage,high processing efficiency and low cost,so it is favored by the vast majority of employers.Online recruitment platform brings people convenient recruitment,but also brings the problem of information explosion to a certain extent.In this paper,through the research of the related technology in the field of artificial intelligence natural language processing,through the relationship between the delivered resume and recruitment information,the matching between resume text and recruitment information text is studied,and a matching algorithm between resume and recruitment information based on text similarity is realized.The text content of resume and recruitment information is divided into structured text and unstructured text for matching calculation,and a matching scheme between resume and recruitment information based on text similarity is designed.Through this solution to solve the problem of low efficiency of recruitment platform matching at this stage.The main contents of this project are as follows:In the matching calculation of structured text between the resume and recruitment information,a preference weight factor is designed to balance the similarity calculation deviation between job seekers and enterprises.The experimental results show that when the preference weight factor is too personal or too corporate,the accuracy will be significantly reduced.By balancing the preference relationship between the both,the resume and recruitment information similarity matching can be more accurate.In unstructured text matching calculation,word2 vec is used to train word vector,and then the semantic information of word vector is trained by Bidirectional Gated Recurrent Unit to solve the problem that Word2 vec is not sensitive to word order and can not use part of speech association information.By adding Attention mechanism,Improve the weight of keywords.From the experimental results,the accuracy of this scheme is significantly better than other comparative experimental schemes.Finally,according to the matching scheme of this project,a bidirectional recruitment matching system based on text similarity between the resume and recruitment information is implemented.The core function modules of the system are the AI job recommendation function of job seekers and the AI resume screening function of recruiters.These two modules are based on the matching scheme in this paper,and show a good matching effect.