Report on the Patent Retrieval Task at NTCIR Workshop 3 Makoto Iwayama , Atsushi Fujii , Noriko Kando , Akihiko Takano Tokyo Institute of Technology / Hitachi Ltd. Institute of Library and Information Science, University of Tsukuba National Institute of Informatics Motivation While a number of commercial patent retrieval systems and services have long been operated, patent retrieval has not been paid much attention in the information retrieval community. One of the reasons is the lack of test collection targeting patent information. Although TREC test collection includes patent documents, the proportion of those documents is signi cantly small. Because patent documents are associated with a number of interesting characteristics from a scienti c point of view, such as document length, document structures, and classi cations, it is important to provide a test collection consisting of patent documents and promote research and development on patent information retrieval. Task description The process of patent retrieval di ers signi cantly depending on the purposes of retrieval, such as technology survey and invalidity search. In NTCIR-3, we focused on a speci c type of technology survey, in which end-users of the retrieval task are novice users, for instance, business managers. We
/lp/association-for-computing-machinery/report-on-the-patent-retrieval-task-at-ntcir-workshop-3-CxYRdJjp5V