• Proceedings of the Korean Society for Information Management Conference
• /
• 1999.08a
• /
• pp.25-28
• /
• 1999

자동분류나 정보검색에 주로 이용되는 문헌 클러스터링에서는 문헌간의 유사성을 측정하기 위해 다양한 유사계수를 이용하는데, 모든 유사계수가 동일한 클러스터링 결과를 가져오는 것은 아니다. 본고에서는 50건의 신문기사를 대상으로 SPSS 통계 패키지를 이용하여 다양한 유사계수에 각각 달라지는 문헌 클러스터링의 결과를 살펴본 후, 유사계수간의 연관성을 측정하였다.

• Journal of the Korean Society for information Management
• /
• v.21 no.1
• /
• pp.93-117
• /
• 2004

This study is to develop a hierarchic clustering model fur document classification and browsing in OPAC systems. Two automatic indexing techniques (with and without controlled terms), two term weighting methods (based on term frequency and binary weight), five similarity coefficients (Dice, Jaccard, Pearson, Cosine, and Squared Euclidean). and three hierarchic clustering algorithms (Between Average Linkage, Within Average Linkage, and Complete Linkage method) were tested on the document collection of 175 books and theses on library and information science. The best document clusters resulted from the Between Average Linkage or Complete Linkage method with Jaccard or Dice coefficient on the automatic indexing with controlled terms in binary vector. The clusters from Between Average Linkage with Jaccard has more likely decimal classification structure.

• Journal of the Korean Society for information Management
• /
• v.18 no.2
• /
• pp.203-230
• /
• 2001

The purpose of this study is to develop a document clustering model for automatic classification of knowledge. Two test collections of newspaper article texts and journal article abstracts are built for the clustering experiment. Various feature reduction criteria as well as term weighting methods are applied to the term sets of the test collections, and cosine and Jaccard coefficients are used as similarity measures. The performances of complete linkage and K-means clustering algorithms are compared using different feature selection methods and various term weights. It was found that complete linkage clustering outperforms K-means algorithm and feature reduction up to almost 10% of the total feature sets does not lower the performance of document clustering to any significant extent.

• Proceedings of the Korean Society for Information Management Conference
• /
• 2000.08a
• /
• pp.45-50
• /
• 2000

자동분류나 정보검색에 활용되는 문헌 클러스터링 결과의 성능을 평가하는 방법에는 여러가지가 있다. 본 논문에서는 제시된 몇 가지 평가방법의 개념과 특징에 대해 알아본다 학술논문 초록 집합인 KTSET과 신문기사 집합인 KFCM-CL을 대상으로 각각 유사계수를 변화시켜가며 클러스터링한 결과에 대해 각 평가방법을 적응해본 후, 특징과 문제점을 살려 보았다.

• Proceedings of the Korean Society for Information Management Conference
• /
• 2005.08a
• /
• pp.65-70
• /
• 2005

계층적 문서 클러스터링에 있어서 실험집단에 따라 응집식 기법과 분할식 기법의 성능이 다르며, 이를 좌우하는 요소는 분류의 깊이, 즉 분류수준이라고 가정하였다. 조금만 나누면 되는 대분류인 경우는 상대적으로 분할식 기법이 유리하고, 조금만 합치면 되는 소분류인 경우에는 응집식 기법이 유리할 것이라고 판단했기 때문이다. 그에 따라 분할식 클러스터링 기법인 양분(Bisecting) K-means기법과 응집식 기법인 완전연결, 평균연결, WARD기법의 성능을 실험집단이 대분류인 경우와 소분류인 경우의 유사계수를 적용하여 각 기법별 성능을 비교하여 실험집단의 특성에 따른 적합 클러스터링 기법을 찾고자 하였다. 실험결과 응집식 기법과 분할식 기법의 성능 우열에 영향을 미치는 것은 분류수준보다는 변이계수로 측정된 상대적인 군집의 크기 편차인 것으로 나타났다.

• Journal of the Korean Society for information Management
• /
• v.23 no.3 s.61
• /
• pp.147-165
• /
• 2006

In this study, experiments for selection of association terms were conducted in order to discover the optimum method in selecting additional terms that are related to an initial query term. Association term sets were generated by using support, confidence, and lift measures of the Apriori algorithm, and also by using the similarity measures such as GSS, Jaccard coefficient, cosine coefficient, and Sokal & Sneath 5, and mutual information. In performance evaluation of term selection methods, precision of association terms as well as the overlap ratio of association terms and relevant documents' indexing terms were used. It was found that Apriori algorithm and GSS achieved the highest level of performances.

• Journal of the Korean Society for Library and Information Science
• /
• v.38 no.1
• /
• pp.35-50
• /
• 2004

This study evaluated the applicability of the static hierarchic clustering model to clustering query results in OPAC. Two clustering methods(Between Average Linkage(BAL) and Complete Linkage(CL)) and two similarity coefficients(Dice and Jaccard) were tested on the query results retrieved from 16 title-based keyword searchings. The precision of optimal dusters was improved more than 100% compared with title-word searching. There was no difference between similarity coefficients but clustering methods in optimal cluster effectiveness. CL method is better in precision ratio but BAL is better in recall ratio at the optimal top-level and bottom-level clusters. However the differences are not significant except higher recall ratio of BAL at the top-level duster. Small number of clusters and long chain of hierarchy for optimal cluster resulted from BAL could not be desirable and efficient.

• Journal of KIISE:Databases
• /
• v.35 no.6
• /
• pp.481-494
• /
• 2008

In this paper we propose Fourier magnitudes based privacy preserving clustering on time-series data. The previous privacy-preserving method, called DFT coefficient method, has a critical problem in privacy-preservation itself since the original time-series data may be reconstructed from privacy-preserved data. In contrast, the proposed DFT magnitude method has an excellent characteristic that reconstructing the original data is almost impossible since it uses only DFT magnitudes except DFT phases. In this paper, we first explain why the reconstruction is easy in the DFT coefficient method, and why it is difficult in the DFT magnitude method. We then propose a notion of distance-order preservation which can be used both in estimating clustering accuracy and in selecting DFT magnitudes. Degree of distance-order preservation means how many time-series preserve their relative distance orders before and after privacy-preserving. Using this degree of distance-order preservation we present greedy strategies for selecting magnitudes in the DFT magnitude method. That is, those greedy strategies select DFT magnitudes to maximize the degree of distance-order preservation, and eventually we can achieve the relatively high clustering accuracy in the DFT magnitude method. Finally, we empirically show that the degree of distance-order preservation is an excellent measure that well reflects the clustering accuracy. In addition, experimental results show that our greedy strategies of the DFT magnitude method are comparable with the DFT coefficient method in the clustering accuracy. These results indicate that, compared with the DFT coefficient method, our DFT magnitude method provides the excellent degree of privacy-preservation as well as the comparable clustering accuracy.

• Journal of KIISE:Software and Applications
• /
• v.34 no.11
• /
• pp.971-983
• /
• 2007

Cellular Genetic Algorithms(CGAs) are a subclass of Genetic Algorithms(GAs) in which each individuals are placed in a given geographical distribution. In general, CGAs# population space is a regular network that has relatively long characteristic path length and high clustering coefficient in the view of the Networks Theory. Long average path length makes the genetic interaction of remote nodes slow. If we have the population#s path length shorter with keeping the high clustering coefficient value, CGAs# population space will converge faster without loss of diversity. In this paper, we propose Smallest-Small-World Cellular Genetic Algorithms(SSWCGAs). In SSWCGAs, each individual lives in a population space that is highly clustered but having shorter characteristic path length, so that the SSWCGAs promote exploration of the search space with no loss of exploitation tendency that comes from being clustered. Some experiments along with four real variable functions and two GA-hard problems show that the SSWCGAs are more effective than SGAs and CGAs.

• Journal of the Korean Society for Library and Information Science
• /
• v.43 no.3
• /
• pp.399-416
• /
• 2009

In this study, clustering methods with related tags were discussed for improving search and exploration in the tag space. The experiments were performed on 10 Delicious tags and the strongly-related tags extracted by each 300 documents, and hierarchical and non-hierarchical clustering methods were carried out based on the tag co-occurrences. To evaluate the experimental results, cluster relevance was measured. Results showed that Ward's method with cosine coefficient, which shows good performance to term clustering, was best performed with consistent clustering tendency. Furthermore, it was analyzed that cluster membership among related tags is based on users' tagging purposes or interest and can disambiguate word sense. Therefore, tag clusters would be helpful for improving search and exploration in the tag space.