We developed a thesaurus for the purpose of natural language processing such as parsing or the term standardization. Because each entry term has a large number of terms with various semantic relations, we introduce a facet and classify them for finding relative terms easily. Furthermore, we distinguish discriminatory terms, and fluctuating Japanese spellings. Thesaurus usage and business development using a thesaurus were also mentioned.

　A full-text search interface connecting to the disaster article database maintained by the authors has condemned users to type a search word in each case. To save the effort of search word selection, we attempted an article retrieval support using word cloud, which displays words in various sizes depending on the occurrence rate. We developed a prototype web page including an SVG image for the word cloud. Users can conduct search by clicking the word on the image, through the JavaScript event handling.

Along with the recent activation of open science, the demand for a repository system that is readable to both humans and machines is increasing. In this context, we constructed data repositories for Institute of Nature and Environmental Technology and Center for Cultural Resource Studies in Kanazawa University using WEKO, which is a repository platform developed by the National Institute of Informatics. In the original design of the WEKO, it is necessary to utilize software for Windows for batch registration of data contents. In the present study, we developed a new user interface for data management for data owners who are not familiar with ICT. By using this system, it is possible for data owners to operate a batch processing for additional registration, update of the registered contents and modication of the metadata simply by uploading a le via browser without depending on the OS. Particularly, it is innovative that out software supports revision control and data access control of the contents by setting "update ag"attribute. In this paper, we introduce the user interface developed for the data repositories in Kanazawa University.

　Several situations such as student's decision making on the application to university and citizen's participation in science communication require to nd interesting academic disciplines. We propose a method for nding academic disciplines using Shinsyo, nonction paperback books. Shinsyo covers several subjects and is easy to read for beginners. We implemented and evaluated two methods of calculating similarity between Shinsyo and academic disciplines. First method calculates BM25 term weights both on Shinsyo and academic disciplines, and then calculates the cosine similarity between them. Second method calculates ranking scores based on BM25 using Shinsyo as a query and academic disciplines as a document. As a result, we showed that the rst method was better than second one.

　“Literary Dojinshi” is Japan’s own media generated as products of literary activities. That media is an important resources for research about Japanese novelist in early stages. Especially, Literary Dojinshi in “Bungaku-Furima” are important resources for contemporary movement of literature in Japan. Nevertheless, there is any comprehensive and systematic databases for Literary Dojinshi in the world. We consider data models for Literary Dojinshi based on ten thousand resources or more deposited in Nihon University College of Arts. As a result, we detected a problem that exist data models about Literary Dojinshi cannot describe relationship of consignment sale in market of Literary Dojinshi.