engleski

Educational Data Driven Decision Making: Early Identification of Students at Risk by Means of Machine Learning

In the last few years there has been a notable increase in the data mining usage for educational purposes. Educational data mining is emerging field of research which has the aim of analysing data about students` activities. Prediction of student achievements is among the fastest growing research in this domain. Main goal of this paper is to provide useful knowledge to faculties and their management using data about students` activity at the LMS Moodle and comparing different machine learning techniques in order to analyse this data. In this paper we have evaluated four machine learning algorithms: neural networks, decision tree, support vector machines and logistic regression. Decision tree shown to be most accurate predictive model. Results indicated lecture and seminar attendance as significant predictors of academic success.

Data-driven educational decision making, decision support system, machine learning, academic performance.

nije evidentirano