• Journal of the Korean Society for information Management
• /
• v.3 no.1
• /
• pp.3-16
• /
• 1986

This research deals with the problems encountered in designing systems for more efficient and effective information retrieval used in the proliferation of literature. This research was designed to develop and test 1) the partitioning a large bibliographic data base into quality oriented subsets (quality filtering), and 2) a system for effective and efficient Information retrieval within subsets of data base (relevance). In order to accomplish this partitioning, the 'kernel' technique of graph theory was applied. In addition, a method of quality filtering utilizing the 'epidemic' theory and the 'obsolescence' of scientific literature was developed.

• Journal of the Korean Society for information Management
• /
• v.8 no.1
• /
• pp.100-116
• /
• 1991

In this study, a retrieval system taking advantage of term relationship in a specific domain and also of evidential reasoning as tools for measuring relevance is implemented. For this experiment, techincal memoranda documented in Electronics and Telecommunications Research Institute (ETRI) served as a sample document file. Sample knowledge base was prepared by extracting terms and term relations pertaining to telecommunications from INSPEC thesaurus. Relations between terms were represented by numerical values according to types of term relations. Relationship between a query and a document was measured according to Dempster-Shafer theory of evidence. As a result of this experiment, a more comprehensive search was made by expanding search terms utilizing term relations. Measure of relevance represented by reflecting term relations, and search results were listed in a descending order of relevance.

• The Transactions of the Korea Information Processing Society
• /
• v.4 no.2
• /
• pp.456-464
• /
• 1997

We describe a random prcess model that prvides sets of patterms whth prcisely contrlolled within-class varia-bility and between-class distinctions.We used these pattems in a simulation study wity the back-propagation netwoek to chracterize its perfotmance as we varied the process-controlling parameters,the statistical differences between the processes,and the random noise on the patterns.Our results indicated that grneralized statistical difference between the processes genrating the patterns provided a good predictor of the difficulty of the clssi-fication problem. Also we analyzed the performance of the Bayes classifier whith the maximum-likeihood cri-terion and we compared the performance of the neural network to that of the Bayes classifier.We found that the performance of neural network was intermediate between that of the simulated and theoretical Bayes classifier.

• Korean Journal of Construction Engineering and Management
• /
• v.1 no.1 s.1
• /
• pp.53-62
• /
• 2000

This study proposes FREES(Fuzzy Risk Evaluation Expert System) for analyzing and evaluating risks occurring during the construction process. The feasibility of this system model was tested by virtual scenario. For the development of the model, at first, risk breakdown structure was established based on risks identified in the existing researches, that is quantitative and qualitative. FREES can reflect human cognition process in the risk analysis and evaluation by adopting artificial intelligence fuzzy theory, differentiating the existing quantitative analysis model. The FREES can be applied to all the project phases from planning to operation & maintenance stage.

• Journal of Digital Contents Society
• /
• v.2 no.2
• /
• pp.103-115
• /
• 2001

This paper proposes the two algorithms which generate a minimal decision rule and approximate inference operation, adapted the rough set and the factor space theory in fuzzy knowledge base. The generation of the minimal decision rule is executed by the data classification technique and reduct applying the correlation analysis and the Bayesian theorem related attribute factors. To retrieve the specific object, this paper proposes the approximate inference method defining the membership function and the combination operation of t-norm in the minimal knowledge base composed of decision rule. We compare the suggested algorithms with the other retrieval theories such as possibility theory, factor space theory, Max-Min, Max-product and Max-average composition operations through the simulation generating the object numbers and the attribute values randomly as the memory size grows. With the result of the comparison, we prove that the suggested algorithm technique is faster than the previous ones to retrieve the object in access time.

• Journal of KIISE:Databases
• /
• v.29 no.5
• /
• pp.347-357
• /
• 2002

Graphs have provided a powerful methodology to solve a lot of real-world problems, and therefore there have been many proposals on the graph representations and algorithms. But, because most of them considered only memory-based graphs, there are still difficulties to apply them to large-scale problems. To cope with the difficulties, this paper proposes a graph representation and graph algorithms based on the well-developed relational database theory. Graphs are represented in the form of relations which can be visualized as relational tables. Each vertex and edge of a graph is represented as a tuple in the tables. Graph algorithms are also defined in terms of relational algebraic operations such as projection, selection, and join. They can be implemented with the database language such as SQL. We also developed a library of basic graph operations for the management of graphs and the development of graph applications. This database approach provides an efficient methodology to deal with very large- scale graphs, and the graph library supports the development of graph applications. Furthermore, it has many advantages such as the concurrent graph sharing among users by virtue of the capability of database.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.40 no.4
• /
• pp.13-20
• /
• 2003

The model of the transistor saturation current of the BJT for integrated circuits based upon the semiconductor physics is proposed. The method for calculating the doping profile in the base region using process conditions is presented and the method for calculating the base Gummel number of lateral PNP BJT and vertical NPN BJT is proposed. The transistor saturation currents of NPN BJT using 20V and 30V process conditions obtained from the proposed method show an average relative error of 6.7% compared with the measured data and the transistor saturation currents of PNP BJT show an average relative error of 6.0% compared with the measured data.

• Journal of KIISE:Computer Systems and Theory
• /
• v.33 no.7
• /
• pp.380-389
• /
• 2006

In this paper, we present the MGrid system and its application for the construction of the Glycoconjugates simulation database called e-Glycoconjugates. The MGrid system is an integrated molecular simulation grid system for computing, databases, and analyses. For e-Glycoconjugates, we have been constructing the simulation database for 2,000 glycan chains and 100 glycoproteins until 2008. In this paper, we present the goal, architecture, and current implementation status of the MGrid system, and e-Glycoconjugates.

• Proceedings of the Korean Information Science Society Conference
• /
• 2011.06c
• /
• pp.307-310
• /
• 2011

최근 멀티미디어 데이터의 급속한 증가는 그를 대상으로 하는 다양한 컴퓨팅 기술의 발전을 가져왔다. 이러한 기술이 인간과의 상호 작용에서 그 양적 범위와 질적 깊이를 더해감에 따라, 멀티미디어 데이터 특히 그 중 가장 대표적이라 할 수 있는 이미지 데이터를 의미적으로 이해할 수 있는 방법의 필요성이 대두되고 있다. 이미지의 의미를 이해하기 위해 저수준(low level)의 시각 정보만을 이용하는 경우 인간과의 상호 작용에서 의미 격차(conceptual gap) 문제가 발생할 수 있다. 이미지 객체의 시각 정보들을 가공해서 온톨로지(ontology)와 같은 형태의 지식 베이스(knowledge base)와 연동하여 보다 고수준의 의미를 부여하는 경우에는 해당 도메인을 벗어난 새로운 환경에 대해 적응력과 강인함이 떨어진다. 이러한 문제를 근본적으로 해결하기 위해서는 지식 베이스가 없는 상태에서 이미지 데이터의 형태로 주어진 대상 객체로부터 의미를 부여할 수 있는 정보들을 추출해, 구조적으로 지식 베이스를 형성해 나가고 이를 토대로 대상 이미지 객체의 의미를 이해할 수 있는 시스템이 필요하다. 본 논문에서는 발달 심리학 이론들을 바탕으로 시각과 관련된 인간의 생득적 능력을 찾고, 이를 기반으로 우선 주어진 이미지 객체로부터 의미 정보를 효과적으로 추출할 수 있는 방법을 제안한다.

• Journal of KIISE:Databases
• /
• v.29 no.5
• /
• pp.367-380
• /
• 2002

Due to recent increase in applications requiring huge amount of data such as spatial data analysis and image analysis, clustering on large databases has been actively studied. In a hierarchical clustering method, a tree representing hierarchical decomposition of the database is first created, and then, used for efficient clustering. Existing hierarchical clustering methods mainly adopted the bottom-up approach, which creates a tree from the bottom to the topmost level of the hierarchy. These bottom-up methods require at least one scan over the entire database in order to build the tree and need to search most nodes of the tree since the clustering algorithm starts from the leaf level. In this paper, we propose a novel top-down hierarchical clustering method that uses multidimensional indexes that are already maintained in most database applications. Generally, multidimensional indexes have the clustering property storing similar objects in the same (or adjacent) data pares. Using this property we can find adjacent objects without calculating distances among them. We first formally define the cluster based on the density of objects. For the definition, we propose the concept of the region contrast partition based on the density of the region. To speed up the clustering algorithm, we use the branch-and-bound algorithm. We propose the bounds and formally prove their correctness. Experimental results show that the proposed method is at least as effective in quality of clustering as BIRCH, a bottom-up hierarchical clustering method, while reducing the number of page accesses by up to 26～187 times depending on the size of the database. As a result, we believe that the proposed method significantly improves the clustering performance in large databases and is practically usable in various database applications.