A methodology for engineering ontology acquisition and validation

  • Authors:

  • Zhanjun Li;Maria c. Yang;Karthik Ramani

  • Affiliations:

  • Alibre incorporated, richardson, texas, usa;Department of mechanical engineering and engineering system division, massachusetts institute of technology, cambridge, massachusetts, usa;School of mechanical engineering, and school of electrical and computer engineering, purdue university, west lafayette, indiana, usa

  • Venue:
  • Artificial Intelligence for Engineering Design, Analysis and Manufacturing
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

When engineering content is created and applied during the product life cycle, it is often stored and forgotten. Current information retrieval approaches based on statistical methods and keyword matching are not effective in understanding the context of engineering content. They are not designed to be directly applicable to the engineering domain. Therefore, engineers have very limited means to harness and reuse past designs. The overall objective of our research is to develop an engineering ontology (EO)-based computational framework to structure unstructured engineering documents and achieve more effective information retrieval. This paper focuses on the method and process to acquire and validate the EO. The main contributions include a new, systematic, and more structured ontology development method assisted by a semiautomatic acquisition tool. This tool is integrated with Protégé ontology editing environment; an engineering lexicon (EL) that represents the associated lexical knowledge of the EO to bridge the gap between the concept space of the ontology and the word space of engineering documents and queries; the first large-scale EO and EL acquired from established knowledge resources for engineering information retrieval; and a comprehensive validation strategy and its implementations to justify the quality of the acquired EO. A search system based on the EO and EL has been developed and tested. The retrieval performance test further justifies the effectiveness of the EO and EL as well as the ontology development method.