Research And Application Of Graduates Employment Prediction Model Based On Machine Learning

In recent years,most colleges and universities in China have expanded their enrollment year by year,and the number of graduates has increased year by year.Based on this social situation,the employment pressure of graduates is also increasing.In order to improve the employment rate of graduates,most colleges and universities attach great importance to the employment guidance work.However,most of the employment guidance work has the problems of lack of pertinence and mere formality generally,which can't benefit the students who really needs help.Many universities have a fixed number of management systems,such as student status information,course information,graduation registration,etc.These systems record all kinds of information of students in an all-round way,but for now these systems are only used for archiving and inquiry at present.If we can make full use of the information contained in these systems and find out the main factors affecting employment,we can predict the employment of graduates and offer the results to colleges and universities.The industry guidance department provides effective suggestions to improve the employment rate of graduates.Firstly,this paper studied the correlation theories about feature selection algorithm and prediction algorithm in detail.Aiming at the application scenario of this paper,i.e.employment forecasting,combined with the research results of domestic and foreign scholars on the employment of graduates,this paper makes an in-depth analysis of the factors affecting the employment of graduates,and summarizes the characteristics of student data: large and complex amount of information,high feature dimension,discrete among attributes,both discrete and continuous characteristics,and redundancy.There are many features.According to the above characteristics,this paper proposes a hybrid feature selection algorithm based on mutual information and weight(HMIGW).Through further analysis,HMIGW feature selection algorithm and XGBoost feature selection algorithm are combined to provide a solution for graduate employment prediction.This algorithm is compared with other decision tree algorithms such as stochastic forest algorithm,and its advantages and disadvantages are evaluated and summarized.Finally,through the operational feedback of the algorithm on the "Practical Teaching Management System" of Shenyang University of Aeronautics and Astronautics,the optimization and improvement direction of the algorithm are put forward,and according to the effect of the algorithm,the decision-making and suggestions for promoting the employment of graduates are put forward for different groups.