Evaluación apriorística de la reusabilidad de los objetos de aprendizaje

Abstract

The use of new technologies in education –e-learning– has changed learning processes providing numerous educational advantages. However, developing quality educational materials for use in these systems is a costly task in terms of both time and resources. The solution to this limitation lies in the reuse of educational materials currently available in different repositories. In this way costs would be reduced while quality would be guaranteed. In an attempt to facilitate this reuse, the main aim of this thesis is to define a model that allows automatic estimation of the reusability of learning objects, using metadata as a source of information. To achieve this goal, we will look other authors’ initiatives regarding the evaluation of learning objects and aimed at improving reusability. Based on these studies, different factors that influence reusability are identified. We go on to determine how they might be evaluated quantitatively using the information contained in the metadata, and to software reusability metrics, that support us in this task. Later an evaluation of the reusability prediction model is carried out, studying the Merlot and eLera repositories. The level of interrelation between the various indicators proposed was analyzed, as were comparatively different ways of aggregating them to provide a single final reusability value. To ensure the effectiveness of this proposal for estimating reusability, calculated reusability data is compared with evaluations and usage data taken from the repositories studied. The main contribution of this proposal is that it allows us to determine the reusability of learning objects automatically using metadata as our source of information. This will help users in the task of searching for reusable educational materials. Also were studied the relationships between the proposed reusability measure, user and expert evaluations, and usage data. Based on this analysis we propose a measure of relevance that integrates all the existing quality indicators. One contribution of the measure of relevance is that it would boost the reliability of recommendations, since it integrates different views of quality. In addition it can be calculated automatically, so its sustainability is assured, allowing all educational materials in repositories to be rated.