A Prediction and Visual Analysis Method for Graduation Destination of Undergraduates Based on LambdaMART Model

Yi Chen, Xiaoran Sun, Wenqiang Wei, Yu Dong, Christy Liang

OAI: oai:igi-global.com:315010 • DOI: 10.4018/IJICTE.315010

Predicting graduation destination can help students determine their learning goals in advance, help faculty optimize curriculum and provide career guidance for students. In this paper, the authors first propose a prediction algorithm for graduation destination of undergraduates based on LambdaMART, called PGDU_LM, which uses Spearman correlation coefficient to analyze the correlation between subjects and graduate destinations and extract characteristic subjects, and uses LambdaMART ranking model to calculate students' propensity scores in different graduate destinations. Second, a visual analysis method for students' course grades and graduation destinations is designed to support users to analyze student data from multiple dimensions. Finally, a prediction and visual analysis system for graduation destination of undergraduates, PGDUvis, is designed and implemented. A case study and user evaluation on this system was conducted using the academic data of students from five majors who graduated from a university during 2016-2020, and the results illustrate the effectiveness of this method.