Employment Analysis And Prediction Research Based On Curriculum Knowledge Network Embedding

As the problem of college graduates’ employment difficulties has become more and more prominent,how to effectively predict the employment situation of graduates has attracted more and more attention from college students and employment guidance departments.To better guide college graduates to choose the appropriate employment development direction is not only an effective use of my country’s human resources,but also a major boost to my country’s education.In current employment forecasts,academic performance,as a key feature,plays a decisive role in employment forecasts.Most studies mainly use grade point average(GPA)and autoencoder to represent academic performance.These two methods can solve the heterogeneity of grades and the sparsity of the grade matrix,but also lose a lot of important information,such as the knowledge behind the curriculum and the links between knowledge,and the student’s knowledge system is closely related to employment.Therefore,this paper proposes a framework to quantify students’ knowledge system,and further explores the impact of students’ knowledge system on employment.The specific work of this paper includes the following two parts:(1)Aiming at the insufficiency of academic performance representation methods in current employment prediction,the MOEN(curriculu M t O knowl Edge mappi Ng)framework is proposed to quantify students’ knowledge system.A course knowledge network that fully taps the knowledge system of each student.(2)Aiming at the problem that the TADW algorithm only measures the importance of words by word frequency,and it is difficult to effectively mine deep semantic information,a prediction algorithm GTADW based on attribute network embedding is proposed,which can effectively represent the structure and attribute characteristics of the network.And combined with the embedding vector and demographic characteristics,it can predict the employment destination of students.Compared with traditional methods,the curriculum knowledge network can capture more information,and this method can provide new ideas for college graduates’ employment prediction problem.