College Student Employment Forecast Based On Campus Big Data

In recent years,the number of applicants for college entrance examinations in our country has set new highs.Correspondingly,major universities and colleges across the country are also constantly implementing enrollment expansion policies.The number of graduates has increased tenfold in ten years.At the same time,the rapid development of education informatization has accumulated massive amounts of educational data in colleges and universities,but in practical applications,the value behind them has not been tapped.In this context,based on campus big data(such as:card information,grade information,employment information,etc.),this paper analyzes and calculates the students' behavior in school,and finally uses machine learning related algorithms to establish a college student employment prediction model.On the one hand,this research can provide decision-making suggestions for relevant departments;on the other hand,it provides employment guidance for college students,and provides early warning and guidance services.The relevant research contents of this article are as follows:(1)Data aggregation and fusion.The data used in this article comes mainly from the campus card and digital education system.First,collect campus big data such as One Card.Then pre-process the relevant data:on the one hand,perform operations such as data cleaning,conversion,integration,specification,and fusion;on the other hand,encrypt student information to protect student privacy.Finally,write the processed data to the database and optimize the related database to improve the ef ficiency of database reading.(2)Extraction and analysis of student behavior characteristics.First of all,the data is divided by week,and basic statistics such as the frequency of student behaviors during the week(such as the number of weekly visits to the library,etc.)are initially counted.Secondly,comprehensively use statistical analysis and deep learning to study the changes in student behavior over time:on the one hand,the slope and breakpoints are used to measure the linear change of student behavior;Measure the temporal patterns of student behavior.Finally,comprehensively use statistical analysis and machine learning to screen out the key features that affect the final employment.(3)Construction of the model.A supervised learning algorithm(random forest)is used to construct a predictive classification model to classify the three states of student employment("unemployed","employed" and "advanced").Experiments show that the prediction model constructed in this paper has an accuracy rate of 70.8%,which has achieved the expected effect.This study combines artificial intelligence technology with campus big data,and builds an intelligent prediction model to make a high-precision prediction of college students' employment.This research further broadens the application field of educational data;at the same time,it also provides some new ideas for relevant researchers.