Enabling the creation of domain-specific reference collections to support text-based information retrieval experiments in the architecture, engineering and construction industries

摘要

The increasing importance of text-based information retrieval (IR) developments in the architecture, engineering, and construction industries (AEC) and the lack of sharable testing resources to support these developments call for an approach that can be used to generate domain-specific reference collections. To address this need, the authors investigated the characteristics of the testing environment in AEC and ways to adapt dominant collection preparation methods for the domain. This paper presents the authors' collection generation approach through the preparation process of the Taiwanese National Center for Research on Earthquake Engineering (NCREE) collection. The collection's Chinese-to-English translation instruments are also discussed as matching semantic/linguistic resources are highly valued in AEC's text-based IR developments. The paper also includes a use case for the NCREE collection to show how a collection generated by the proposed approach could be applied to support research experiment and validation. The direct outputs, the NCREE collection and its translation instruments, are sharable and reusable testing resources, while mechanisms for seeking collections from other researchers are part of the extended research endeavors.