Ontological interoperability of learning objects: a hybrid graphical-neural approach

Abstract

This paper presents OntoShare, an automated ontology mapping and merging architecture for learning object retrieval and reuse. The architecture aims to offer contextual and robust ontology mapping and merging through hybrid unsupervised clustering techniques comprising of Formal Concept Analysis (FCA), Self-Organizing Map (SOM) and K-Means clustering incorporated with linguistic processing using WordNet. The merged ontology facilitates sharing and retrieval of learning objects from the Web or from different learning object repositories such as ARIADNE and Educause. Experimental results can be extended to other resources in databases or data warehouses.