Asma Aldrees and Azeddine Chikh and Jawad Berri, King Saud University, Kingdom of Saudi Arabia

The classification of learning objects (LOs) enables users to search for, access, and reuse them as needed. It makes e-learning as effective and efficient as possible. In this article the multilabel learning approach is represented for classifying and ranking multi-labelled LOs, whereas each LO might be associated with multiple labels as opposed to a single-label approach. A comprehensive overview of the common fundamental multi-label classification algorithms and metrics will be discussed. In this article, a new multi-labelled LOs dataset will be created and extracted from ARIADNE Learning Object Repository. We experimentally train four effective multi-label classifiers on the created LOs dataset and then, assess their performance based on the results of 16 evaluation metrics. The result of this article will answer the question of: what is the best multi-label classification algorithm for classifying multi-labelled LOs?

Learning object, data mining, machine learning, multi-label classification, label ranking.