Incorporating Domain Knowledge and Information Retrieval Techniques to Develop an Architectural/Engineering/Construction Online Product Search Engine

Abstract

This paper introduces a domain-specific search engine, which was developed to take advantage of the growing online product information for surveying the virtual product market. Knowledge about a product was uniquely incorporated with query expansion operations and the extended Boolean model retrieval approach to handle issues associated with the search engine development. The search engine was designed to (1) represent and utilize the represented knowledge in the product domain; (2) identify online product information; and (3) then evaluate the collected online product information. A prototype search engine for testing was developed and statistically validated with five data sets with each data set being derived from a different type of product according to MasterFormat (Alexandria, Va.) categorization. The validation results indicated that compared with the tested general search engine or aggregated information service, the prototype was able to identify more distinct product manufacturers for procurement-related decision support.