Background knowledge driven ontology discovery

We have previously proposed a GSOM-based hybrid model for automatic discovery of ontology, the first step towards semi-automation of ontology construction. One of the shortcomings of this previous model is the use of a threshold for selecting abstraction levels. The threshold might introduce an inappropriate concept and cause information loss. In this paper, we introduce a new parameter called context ratio (cr) to overcome this drawback. The cr is used as stopping criteria for traversing hypernyms and identifying appropriate abstraction levels. It allows us to extend the previously proposed framework to integrate methodology for multiple inheritance validation in the discovered ontology.