Metadata Generation for Learning Objects: An Experimental Comparison of Automatic and Collaborative Solutions

Abstract

This paper deals with automatic and collaborative metadata generation for learning objects. Firstly, content- and context-based methods of automatic metadata generation are analyzed with respect to their suitability for generating the metadata types defined in the IEEE LOM, based on the results of an experiment comparing key phrase extraction tools. Secondly, collaborative tagging is discussed as an approach for metadata generation that aims to profit from a possibly large number of users. Potentials and challenges are discussed with the help of an experiment on convergence, commitment and coordination in collaborative tagging. Finally, recommendations are given which motivate for a hybrid approach embracing both, automatic and collaborative solutions to metadata generation.