The knowledge for the specific object is expressed by a way of various explanations. In this study, we took out those explanations about terms from a Japanese-language dictionary and a biology textbook and categorize those explanations by the ways. As a result, the 20 types of the way of expression are extracted. Furthermore, we examined the appearance numbers those types in a Japanese-language dictionary and a biology textbook and clarified that there was deflection among the explanation type.

View full abstract

Ancient documents and scriptures provide a clue as to their background and the human relationships, due to the colophons and the bibliographic data. Using the full-text search system that we have developed, we attempt to extract the helpful keywords. The words of Japanese year are made of the imperial era names and the numbers to filter through the full-text search. Temple names are obtained by a text-mining approach whose first step is to find the Chinese character of temple. The keywords are split character by character and formed by means of trie to store in a database. Moreover we improve the search interface so that the users can see the resulting documents with the keywords highlighted for further search.

View full abstract

There is currently a variety of computerized historical and cultural resource information stored on digital networks, namely the Internet. This information contains a variety of information on historical figures and through the provision of such reference material it is possible to promote a further understanding of historical and cultural archives and documentation. This research considers the potential uses of authentic data concerning historical figures in the modern age that can be shared over digital networks following examination of the current state of such information available over the Internet.

View full abstract

Recently, various fields of science are becoming increasingly specialization. There is a necessity for performing the research in a lot of fields with collaboration. This paper proposes a recommendation system based on Collaborative Filltering for automatically obtain a research partner from the network. By using data from a bibliographic database, this network is constructed in which the nodes are researchers and two researchers are connected if they have co-authored a paper. The proposed system estimates the usefulness of unknown co-authored researcher using Collaborative Filtering algorithms. This suggested the proposed system can be used for discovering useful researchers.

View full abstract

Bibliometrics, the study to quantitatively analyze article information in the database, is quite frequently limited in the no. of analyzable cases due to a manual operation. Japan Science and Technology Agency has newly developed a visualized-analysis technique application capable of analyzing by far a greater no. of cases, identifying the organization names and making a co-appearance map with multiple items. The present paper considers an impact that visualized-analysis technique has on bibliometrics.

View full abstract

Scholarly journals published by Japanese academic societies play important roles in dissemination and sharing of Japanese researchers' research results. To further promote their international distribution, internationalization of the Japanese academic journals in various aspects is required. In recent years, articles submitted from abroad to those journals seem to be increasing. In this study, we look the situation surrounding academic journals in Japan through detailed analysis on those journals with high percentages of foreign researchers' papers, based on the CJP database, produced by National Institute of Informatics.

View full abstract

Collaborating research is made abundantly in the academic research. A difference of aspect by the research field appears especially in collaboration work with several organizations. We develop software for visualization of collaborating research on CJP database and KAKEN database. This software is most useful for analyzing an aspect of fields, trends, and locality of academic research.

View full abstract

The author has been publicizing statistical indexes for Japanese universities based on the number of papers and citation counts in the ISI's databases. Increasing financial difficulty at the national government tends to encourage arguments that some performance-based system should be introduced in allocating funds for national and private universities. In this paper, the author makes a trial calculation to allocate funds according to the citations gained by each university, and contrasts them with the present allocations. Then, yearly progression of the system's effect is traced with a simulation. Discussions on influences and problems caused by this type of meritocracy follow.

View full abstract

When we see the building process of knowledges in science. We can find two kinds of knowledge-building process based on different type of reasoning. The one is "penetrating knowledge process", which is penetrating a special domain knowledge into the object to be understood. The other is "connecting knowledge process", which is connecting plural special domain knowledges to a broader domain knowledge. The type of reasoning in "penetrating knowledge process" is necessary and inevitable analytic-reasoning about the object, using a special domain knowledge. The type of reasoning in "connecting knowledge process" is probable and composing synthetic-reasoning between different domain knowledges. For human survival value, we have to give the same value to both researching activities of two knowledge-building process. We have to research to make up such construction of knowledges, as hamonizes analysis and synthesis of the world. That is the construction of "FUSIONOLOGY" of sciences.

View full abstract

In recent years Kansei retrieval systems have been developed to be practical use. We have developed the retrieval system using Kansei parameters based on online book reviews of children's books. The relation between parameters assigned manually and the words in online book-reviews has yet to be analyzed enough, however. Moreover, how to set appropriate value of Kansei parameters based on the words in online book reviews has not been theoretically proved. To resolve such problem, we have conducted an experiment. In this experiment, Kansei parameters which are consistent with the value gained manually were assigned based on "the words, which represent human feelings, impressions or atmosphere of books" using the machine learning approach. We have used 1605 book reviews to assign the parameters automatically. The result showed that : 1)using the words in book reviews is useful to assign parameters automatically, 2)using about 30 frequent words is more effective than using all the words in book reviews, and 3) the recall ratio is not influenced by the parts of speech of the words extracted from book reviews.

View full abstract

Digital library is one of the library services that provide information resources and searching tools. University of Tsukuba Library provides various kinds of tools, especially "MetaLib" and "SFX" are powerful tools. MetaLib is a metasearch engine, which conduct simultaneous searches in multiple, and often heterogeneous information resources. SFX is a linking system, which allows context-sensitive linking between Web resources. In this paper, we report a trend chart of utilization of digital libraries based on access log analysis. Moreover, we made a tutorial of the digital library system of University of Tsukuba, and evaluated the effect. It was found that the numbers of utilization of the system are increasing. The tutorial did not develop new users, but led existing users to navigate to other effective services.

View full abstract

We constructed a dictionary reading system using relationships between words. Whenever any words are looked up in our system, related words of seven classes are provided, such as synonyms, antonyms, words co-occured in well-known phrases, and so on. Consequently, users can read the dictionary through the network of interesting words. In the preriminary experiment, we found that unknown words and unknown relationships between words were important to make dictionaries attractive reading.

View full abstract

The effect of the utilization of ontology is examined in order to extract causal relationships from scientific papers. At first, engineering ontology was proposed by integrating existing conceptual models about artifacts, and its validity was tested. Then, an experiment to extract causal relationships was performed by using the obtained ontology. The precision in the extraction is significantly improved compared to the case using co-occurrence of terms from 28.6% to 63.3%.

View full abstract

The authors have conducted a user study for 11 undergraduate students in order to understand information seeking behaviors on Web. The participants were assigned two types of information seeking tasks, Informational task and Transactional task, where each task takes 15 minutes. This paper reports preliminary analyses of eye movement data during conducted tasks. Two types of the analysis are shown. One is an analysis based on manual coding for area of interests. The other is based on fixation analysis of rough block assignments into some of display areas.

View full abstract

Clustering using cocitation and bibliographic coupling is an effective tool to analyze the structure of scientific research, but its arbitrariness in the setting of clustering threshold is a problem. This study tracked modularity of citation networks in research domains, and we found that there are three stages in clustering of citation networks and it is universal across our case studies. In the first stage, core clusters in the domain are formed. In the second stage, peripheral clusters are formed, while core clusters continue to grow. In the third stage, core clusters grow again. By focusing on elementary process in clustering, we can understand a hierarchical structure of domain and judge the location of each paper, which cannot be seen in the final clustering results.

View full abstract

We tried to construct the solution supported system for problems in materials field using semantic Web technologies and pointed out necessary functions. In this report, mechanical properties data and model formulas for analysis of data were collected in order to resolute mechanics problems. And then the application of the MathML, that was XML description in the formula to the system was examined. Consequently, in order to solute smoothly data analysis and the problem in materials field was enabled by communalizing unification of the character and symbol which are used with the formulas that were expressed by MathML. Thus, the unification were proposed and an example of the application was introduced.

View full abstract

Based on the government plan "Intellectual Property Strategic Program 2006",We,Japan Science and Technology Agency (JST) has constructed a database "JSTPatM", integrating technological part of patent and science & technology information. We started its free providing service to Universities from March 2007. The paper reports ideas and efforts in data structure, interface and links to patent documents of JSTPatM.

View full abstract

Journal publication in the humanity and social science fields was studied. There are 1436 scholarly journals and 4156 university bulletins are published, and among them, 279 scholarly journals and 2236 university bulletins are digitized and available from NII-ELS. In addition, 37 scholarly journals and 64 university bulletins are on the web and linked from WARP of the National Diet Library

View full abstract