Research On Key Technologies Of Person-post Matching Analysis Based On Big Behavior Data

In recent years,the employment channels of college students have become more and more abundant,and there are more and more categories.However,the recruitment process has not changed substantially and the recruitment cost is high.Moreover,there is a gap between the training and recruitment needs of colleges and universities.To recruit a qualified candidate for a post,a large number of written examinations and interviews are needed,which leads to inefficient recruitment,and the whole process of recruitment is influenced by subjective factors.How to find an efficient and objective solution is very meaningful.With the popularization of campus digitalization,digital campus information system has been accumulating various kinds of information on campus,and forms are becoming more and more diversified,which hides many valuable information and rules.It is necessary to explore the internal relationship between enterprise posts and the comprehensive quality of employed students,and on this basis,to construct a matching analysis mechanism based on student portraits and job demand information.The main contributions include:(1)According to the requirement characteristics of Person-Post matching analysis,the overall design of student portrait model is given,including the definition and quantification of multi-dimensional labels in student portraits.On this basis,a scheme of collecting and pretreatment of students' behavior data in school is put forward.(2)Aiming at the evaluation needs of subjective labels in student portraits,a method of generating rating labels is proposed.An improved SAGA-FCM algorithm based on simulated annealing genetic optimization(SAGA-FCM)is used to cluster the sample data to get the initial classification of rating tags.(3)Based on the clustering results of SAGA-FCM,in order to evaluate the new samples,the improved generalized regression neural network algorithm based on genetic optimization(GA-GRNN)is used to predict the rating tags of new samples.(4)A joint embedding CNN(JE-CNN)used for Person-Post matching is designed and implemented.Through text representation of student portraits and job demand information,the two are transformed into representation vectors in public space,and then the matching degree between students and jobs is analyzed by means of quantitative analysis.(5)The SAGA-FCM,GA-GRNN and JE-CNN are tested and the results are analyzed.The experimental results show that the solution proposed in this paper is effective and can give meaningful guidance and influence to campus recruitment,student self-evaluation and university personnel training.