AUTOMATIC ANNOTATION AND RETRIEVAL SYSTEM (ILARS) FOR ENHANCING ORGANIZATIONAL E-LEARNING.
Keywords:
Learning object,, Metadata annotation, Organizational e-Learning, WordNet, Natural Language Processing (NLP)Abstract
Context independent, reusable learning objects (RLOs) not only allow for easy access to tailored learning but also provide unprecedented efficiency in the construction of learning environments. However, creating RLOs entails certain requirements: technically it requires de-contextualization for increased reusability, while pedagogically it requires context-preservation to provide coherent learning experiences. Thus creating Leaning Objects (LOs) that are both reusable and contextualized can be a difficult challenge. Extending the reusability of LOsto satisfy user needs in a given domain can be achieved by semantically annotatingLO metadata with specific contextual informationrelated to the particular organization. However, given that “context is the friend of learning and the enemy of reuse”, adding such information would subsequently reduce the reusability of the LOs. In this paper, we propose a lexical similarity approach which not only increases the reusability of existing LOs but also relieves LO developers from tedious annotating metadata. The proposed approach optimizes automatic categorization and annotation of LOs in a given domain, thus increasing the efficiency with which learning environments can be developed.