Learning search keywords for construction procurement

Abstract

Seeking information from websites has become an essential part of a contractor's procurement undertaking. as more and more procurement websites become available on the Internet, Websites host extremely lame amounts of information: a keyword search, therefore, is often more efficient than browsing via an index. However. in order to find the desired information. it may be necessary to enter keywords using a trial-and-error process. This research recognizes that professional procurement experience can help users search website information more effectively, by using fewer keywords. and so proposes a learning model and suggestion model that can capture such experience. thus guiding inexperienced users in their search. Experiments, evaluating the performance of the system, were also conducted. (C) 2004 Elsevier B.V. All rights reserved.