The Journal of Information Science and Technology Association
(Johono Kagaku to Gijutsu)
Abstract

Vol. 60 (2010), No.6

Data mining methodologies and application for libraries

On the characteristics and application of data mining

Yoshihiro SUGITA
Gakushuin University, Faculty of Economics
1-5-1 Mejiro, Toshima-ku, Tokyo 171-8588 JAPAN

Abstract: What is data mining? This paper identifies the characteristics of data mining by comparing data mining and traditional statistical analysis. Under today's environment where a large amount of data is readily available, the theory of statistical inference has small relevance and it is possible to perform many data analysis for real world problems at a reasonable cost. Data mining is a methodology developed in such environment. Because it is necessary to understand the process of data mining in order to solve real problems, this paper attempts to explain each step in the process and identifies important concepts and methods.

Keywords:  data mining / data mining process / data mining methods / marketing data analysis / POS data

Utilization of citation information by link mining

Yuya KAJIKAWA
Graduate School of Engineering, The University of Tokyo
2-11-16 Yayoi Bunkyo-ku, Tokyo 113-8656 JAPAN

Abstract: Link mining and text mining are especially important technique for librarian and engineers in library and information center where documents such as books, articles, and patents are transacted. In this survey, methodology, examples, and limitations in utilization of citation information by link mining are illustrated. Topics include application cases like illustrating academic landscape, detection of emerging domains, prediction of future core papers, and extraction of difference among corpus. It also includes the limitations of the analysis such as validity of corpus, selection of link creating method, time lag and granularity of analysis.

Keywords:  data mining / link mining / citation analysis / network analysis / ranking / clustering / structure prediction / academic landscap

Application of text mining.

Minoru YOSHIDA, Hiroshi NAKAGAWA
Information Technology Center, University of Tokyo
General Library 4F, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033 JAPAN

Abstract: Text mining is a task to apply data mining techniques to text data. In typical text mining approaches, data mining techniques are applied to some statistical data derived from text by using natural language processing technologies. In this article, we introduce what is text mining and natural language processing technologies needed for text mining. We also introduce our query support systems UT-Kiwi (for searching documents on the University of Tokyo Web site) and Wikiwi (for searching Wikipedia articles) as examples of text mining applications.

Keywords:  text mining / natural language processing / information retrieval / query support / suffix arrays

Development of a recommendation system to promote of utilization of libraries

Tetsuro AOTO
Network Service Section Center for Science Information, Shibaura Institute of Technology
307 Fukasaku, Minuma-ku, Saitama-city, Saitama 337-8570 JAPAN

Abstract: In late years various information are provided on Internet, which causes the decrease of the students learning by themselves using correct information.
In addition, there isn't enough information which book is necessary for their classes. These are also the reasons the decrease of the users of libraries.
Therefore, we decided to develop a new book-recommendation system to increase library users.
In this paper, we report the developed system which recommend books for students based on the information of the syllabus which are familiar to them, and evaluate it.

Keywords:  TermExtract / N-gram / syllabus / recommendation

Library marketing for improving patron satisfaction
−Expectation to the improvement of library services by data analysis−

Toshiro MINAMI
Kyushu Institute of Information Sciences & Kyushu University Library
6-3-1 Saifu, Dazaifu, Fukuoka 818-0117 JAPAN

Abstract: As was described in the Five Laws of Library Science, libraries are supposed to keep changing and providing their patrons with up-to-date information services. We are in a rapidly changing society now due to the development of mobile network terminals like smart phones. In order to adapt to such a rapid change, libraries should adopt the concept and technologies of marketing. It is the only way for them to survive so that they can keep getting to be accepted as reliable and necessary organization by their patrons. In this paper we define, present potentiality of library markething, and demonstrate its importance. We have to keep investigating the beneficial methods and to construct the service model for the futute libraries.

Keywords:  data analysis / data mining / service marketing / library marketing / e-Society / patron / user-oriented

Page Top Copyright (C) 2010 INFOSTA. All rights reserved.