Learning objects repository management using an adaptive quality evaluation Multi-Agent System

Abstract

Availability and correspondence with expectations are desired characteristics in order to guarantee the quality of Learning Objects (LOs) retrieved from LO repositories during the search process. The administrators of these repositories have the responsibility of ensuring the quality of LOs after applying their corresponding evaluation. The implementation of metrics applied on relevant characteristics of LOs is a crucial tool for LO evaluation. This paper proposes an approach that uses a Multi-Agent System (MAS) for assessing main LO characteristics, applying different methods and metrics being adjustable to different kinds of repositories by employing adaptive parser agents. By using metadata as main source of information, the metrics allow users to rate the quality of LOs and generates alarms concerning inputs that do not meet the expected values.The system developed automatically evaluates a large number of resources to facilitate the work of the repository administrators before improving or publishing the LOs into a repository federation. © Springer-Verlag Berlin Heidelberg 2013.