论文信息 - Using GitLab Interactions To Predict Student Success When Working As Part Of A Team

Using GitLab Interactions To Predict Student Success When Working As Part Of A Team

This paper explores machine learning algorithms that can be used to predict student results in an assignment of a Software Engineering course, based on weekly cumulative average source code submissions to GitLab. GitLab is a source code version control system, commonly used in Software Engineering courses in Higher Education. The aim of this work is to create models that can be used to predict if a group of students in a team will pass or fail an assignment. In this pa-per, we present results from Decision Tree, Random Forest, Extra Trees, Ada Boost and Gradient Boosting machine learning models. These models were evaluated using cross-validation, with Ada Boost achieving the highest average score.

John Domingue | Alexander Mikroyannidis | Allan Third | Audrey Ekuban

[1] Vicente Matellán Olivera,et al. Model for Evaluating Student Performance Through Their Interaction With Version Control Systems , 2018, LASI-SPAIN.

[2] George Siemens,et al. Learning analytics: envisioning a research discipline and a domain of practice , 2012, LAK.

[3] Dragan Gašević,et al. Let's shine together!: a comparative study between learning analytics and educational data mining , 2020, LAK.

[4] Ryan S. Baker,et al. The State of Educational Data Mining in 2009: A Review and Future Visions. , 2009, EDM 2009.

[5] J. Hanley,et al. A method of comparing the areas under receiver operating characteristic curves derived from the same cases. , 1983, Radiology.

[6] Gaël Varoquaux,et al. Scikit-learn: Machine Learning in Python , 2011, J. Mach. Learn. Res..

[7] Caroline Jay,et al. A Methodology for Using GitLab for Software Engineering Learning Analytics , 2019, 2019 IEEE/ACM 12th International Workshop on Cooperative and Human Aspects of Software Engineering (CHASE).