Are Collaborative Filtering Methods Suitable for Student Performance Prediction?
|Year of publication||2015|
|Type||Article in Proceedings|
|Conference||Progress in Artificial Intelligence - 17th Portuguese Conference on Artificial Inteligence - EPIA 2015|
|MU Faculty or unit|
|Keywords||Student Performance; Prediction; Collaborative Filtering Methods; Recommender System|
|Description||Researchers have been focusing on prediction of students’ behavior for many years. Different systems take advantages of such revealed information and try to attract, motivate, and help students to improve their knowledge. Our goal is to predict student performance in particular courses at the beginning of the semester based on the student’s history. Our approach is based on the idea of representing students’ knowledge as a set of grades of their passed courses and finding the most similar students. Collaborative filtering methods were utilized for this task and the results were verified on the historical data originated from the Information System of Masaryk University. The results show that this approach is similarly effective as the commonly used machine learning methods like Support Vector Machines.|