Job Recommendation For Graduates Based On Multiple Social Relationships And Migration Patterns

Recently,as the number of college graduates has increased year by year,the overall competition in the job market in China has shown an intensified trend,and college graduates without work experience are facing severe employment problems.The graduate job recommendation system is to recommend some of the most suitable jobs for graduates,avoiding the blind application of graduates due to lack of employment experience,saving graduates’ time and energy spent in the employment process,thereby increasing the success rate of graduates’ employment and employment satisfaction,therefore,graduate employment recommendation is an important research direction in the recommendation system.Employment recommendation for graduates is a challenging problem.In the past,the recommendation results obtained by using traditional recommendation algorithms were generally not good.The main reason for this phenomenon is that most graduates have no employment history before they leave the campus.The traditional collaborative filtering recommendation method has the problems of data sparseness and user cold start.With the acceleration of the modernization process in education,intelligent education management systems and student behavior monitoring systems have been widely used,and the resulting educational big data has also been increasingly applied to students’ academic and behavioral predictions.This article mainly uses the past employment records of students provide employment recommendations for graduates.In order to explore the connections between students and their respective employment preferences,this research builds a graduate employment recommendation system based on a large amount of personal information,academic and behavioral data generated by students in campus life.The main research contents of this paper are as follows:First of all,the current employment recommendation system algorithm and the current research status of education data mining technology are expounded and analyzed,and the algorithm is briefly introduced,focusing on the specific algorithms of graph neural network and recurrent neural network used in current deep learning.Secondly,in order to explore the effective value of various student-related information in campus big data,first introduces the fields and meanings of various data in campus big data,and uses statistical analysis to study various factors that affect students’ employment choices.Third,this article finds that the performance of graduates’ employment recommendation can be improved by considering the social relationships between students and their migration patterns between cities.A work recommendation algorithm(MSR-MP)based on multi-graph network and recurrent neural network considering multiple social relations and migration patterns is proposed.The attention mechanism is used in the multi-graph network to capture the employment preferences of users caused by the multiple social influences between graduates,and the recurrent neural network can model the migration pattern of graduates between cities to predict the graduate’s work location.Finally,the experimental results on the real graduate employment data set prove that the method proposed in this study is better than the recommendation effect of several baseline models.When the length of the job recommendation list is 20,the MSR-MP hit rate of the model(HR)reached 57.03%,which is about twice the Bayesian personalized ranking algorithm(BPR),and the average ranking reciprocal(MRR)reached 35.74%,which is higher than the college student employment recommendation algorithm based on personalized preference(P2CF)19 percentage points.