TY - GEN
T1 - Intelligent search guides for the construction e-market
AU - Dzeng, Ren-Jye
AU - Wang, S. S.
AU - Chang, S. Y.
PY - 2005/12/1
Y1 - 2005/12/1
N2 - As information floods the Internet and the complexity of projects increases, efficient searches for procurement information on the Internet have become essential for contractors. A procurement engineer as a buyer must be both experienced in procuring items and familiar with the content and terms used on a construction procurement website or Internet, to be able efficiently to complete procurement documents. A buyer finds desired procurement information by browsing data using menus or indexes pre-designed on a website, or by inputting search keywords. Using keywords to seek information is frequently more efficient when a website hosts a large amount of information, or the buyer does not know the specific names of procured items used by the website. Nevertheless, people who are unfamiliar with the website may have to perform a trial-and-error process of inputting keywords, even though they may be professionally experienced in procurement. This work examines the search behaviors of procurement engineers, and presents a guide system that helps engineers reduce the trial-and-error process through a learning framework, which elicits knowledge about the search patterns of procurement experts. When an engineer inputs a keyword phrase, the system either replaces the keyword with a correct or more specific one, or adds other keywords that target related procurement items. An evaluation that involves 12 professionals and 64 procurement items reveals that the proposed guide system enables a dynamic and customized guide menu, and reduces the number of keywords required to find the desired information.
AB - As information floods the Internet and the complexity of projects increases, efficient searches for procurement information on the Internet have become essential for contractors. A procurement engineer as a buyer must be both experienced in procuring items and familiar with the content and terms used on a construction procurement website or Internet, to be able efficiently to complete procurement documents. A buyer finds desired procurement information by browsing data using menus or indexes pre-designed on a website, or by inputting search keywords. Using keywords to seek information is frequently more efficient when a website hosts a large amount of information, or the buyer does not know the specific names of procured items used by the website. Nevertheless, people who are unfamiliar with the website may have to perform a trial-and-error process of inputting keywords, even though they may be professionally experienced in procurement. This work examines the search behaviors of procurement engineers, and presents a guide system that helps engineers reduce the trial-and-error process through a learning framework, which elicits knowledge about the search patterns of procurement experts. When an engineer inputs a keyword phrase, the system either replaces the keyword with a correct or more specific one, or adds other keywords that target related procurement items. An evaluation that involves 12 professionals and 64 procurement items reveals that the proposed guide system enables a dynamic and customized guide menu, and reduces the number of keywords required to find the desired information.
KW - Construction procurement
KW - Ecommerce
KW - Information search
KW - Knowledge acquisition
KW - machine learning
UR - http://www.scopus.com/inward/record.url?scp=80053418335&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:80053418335
SN - 1905088000
SN - 9781905088003
T3 - Proceedings of the 10th International Conference on Civil, Structural and Environmental Engineering Computing, Civil-Comp 2005
BT - Proceedings of the 10th International Conference on Civil, Structural and Environmental Engineering Computing, Civil-Comp 2005
Y2 - 30 August 2005 through 2 September 2005
ER -