초록

The purpose of this study is to generate the local level knowledge structure of a single document, similar to end-of-the-book indexes and table of contents of printed material, through the use of term clustering and cluster representative term selection. Furthermore, it aims to analyze the functionalities of the knowledge structure, and to confirm the applicability of these methods in user-friendly information services. The results of the term clustering experiment showed that the performance of the Ward's method was superior to that of the fuzzy K-means clustering method. In the cluster representative term selection experiment, using the highest passage frequency term as the representative yielded the best performance. Finally, the result of user task-based functionality tests illustrate that the automatically generated knowledge structure in this study functions similarly to the local level knowledge structure presented in printed material.攀＊＊＊ 본 연구는 연세대학교 대학원 박사학위논문의 일부를 요약한 것임.＊＊＊ 日本 慶應義塾大學(Keio University) 圖書館情報學科 訪問硏究員(libinfo@yonsei.ac.kr)＊＊＊＊연세대학교 문헌정보학과 교수(ymchung@yonsei.ac.kr) 논문접수일자 : 2004년 8월 17일 게재확정일자 : 2004년 9월 10일攀攀

Abstract

초록

Abstract

Titles have been regarded as having effective clustering features, but they sometimes fail to represent the topic of a document and result in poorly generated document clusters. This study aims to improve the performance of document clustering with titles by suggesting titles in the citation bibliography as a clustering feature. Titles of original literature, titles in the citation bibliography, and an aggregation of both titles were adapted to measure the performance of clustering. Each feature was combined with three hierarchical clustering methods, within group average linkage, complete linkage, and Ward's method in the clustering experiment. The best practice case of this experiment was clustering document with features from both titles by within-groups average method.