A domain ontology model for mould design automation

Abstract

An ontology-based search model with semantic distance measures is proposed to improve the traditional keyword-based search for the mould design domain. The model has three components. First an NLP component is used to extract independent concepts from text with keywords extracted from sentences. Next, the ontology layer is built to process concepts with minimal total semantic distance to all these keywords found with a ranking algorithm. Finally, the concepts in the ontology are mapped to the concepts in a proprietary database to implement the matching process from sentences to database concepts; enabling integration with existing mould design software. The ontological search is compared against traditional keyword based search in the mould design domain and showed more fault tolerance and flexibility in maximizing the accuracy and number of detected matches. © 2010 Springer-Verlag Berlin Heidelberg.