• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2002.05a
• /
• pp.119-122
• /
• 2002

The research Interest of this paper is focused on the efficient clustering task for an arbitrary color data. In order to tackle this problem, we have tiled to model the inherent uncertainty and vagueness of color data using fuzzy color model. By laking a fuzzy approach to color modeling, we could make a soft decision for the vague regions between neighboring colors. The proposed fuzzy color model defined a three dimensional fuzzy color ball and color membership computation method with the two inter-color distance measures. With the fuzzy color model, we developed a new fuzzy clustering algorithm for an efficient partition of color data. Each fuzzy cluster set has a cluster prototype which is represented by fuzzy color centroid.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.5 no.2
• /
• pp.58-69
• /
• 1995

In this paper GFI (Generalized Fuzzy Isodata) and FI (Fuzzy Isodata) algorithms are studied and applied to the tire tread pattern classification problem. GFI algorithm which repeatedly grouping the partitioned cluster depending on the fuzzy partition matrix is general form of GI algorithm. In the constructing the binary tree using GFI algorithm cluster validity, namely, whether partitioned cluster is feasible or not is checked and construction of the binary tree is obtained by FDH clustering algorithm. These algorithms show the good performance in selecting the prototypes of each patterns and classifying patterns. Directions of edge in the preprocessed image of tire tread pattern are selected as features of pattern. These features are thought to have useful information which well represents the characteristics of patterns.

• Journal of Korea Society of Digital Industry and Information Management
• /
• v.5 no.1
• /
• pp.23-32
• /
• 2009

Although the web is becoming a widely accepted medium, it provides relatively poor performance and low availability. A cluster consists of a collection of interconnected stand-alone computers working together and provides a high-availability solution in application area such as web services or information systems. Web server clusters require a high-availability service with a proactive and practical fault management. However, as the system complexity grows, it is not easy to meet the requirement. Therefore, web server clusters must have self-fault management capability for meeting high-availability requirement. In this paper, we propose high availability web server clusters using self-healing technique with a minimal human intervention. Our experimental results show that a proposed method can be used to improve the availability of web server clusters.

• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 1996.10a
• /
• pp.31-34
• /
• 1996

The extensive researches have compared the performance of neural networks(NN) with those of various statistical techniques for the classification problem. The empirical results of these comparative studies have indicated that the neural networks often outperform the traditional statistical techniques. Moreover, there are some efforts that try to combine various classification methods, especially multivariate discriminant analysis with neural networks. While these efforts improve the performance, there exists a problem violating robust assumptions of multivariate discriminant analysis that are multivariate normality of the independent variables and equality of variance-covariance matrices in each of the groups. On the contrary, cluster analysis alleviates this assumption like neural networks. We propose a new approach to classification problems by combining the cluster analysis with neural networks. The resulting predictions of the composite model are more accurate than each individual technique.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.14 no.3
• /
• pp.369-374
• /
• 2004

A cluster merging algorithm that merges convex clusters resulted by the Fuzzy Convex Clustering(FCC) method into non-convex clusters was proposed. This was achieved by proposing a fast and reliable distance measure between two convex clusters using Support Vector Machines(SVM) to improve accuracy and speed over other existing conventional methods. In doing so, it was possible to reduce cluster number without losing its representation of the data. In this paper, results for several data sets are given to show the validity of our distance measure and algorithm.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.11 no.7
• /
• pp.579-583
• /
• 2001

In the area of data clustering in high dimensional space, one of the difficulties is the time-consuming process for computing vector similarities. It becomes worse in the case of the agglomerative algorithm with the group-average link and mean centroid method, because the cluster similarity must be recomputed whenever the cluster center moves after the merging step. As a solution of this problem, we present an incremental method of similarity computation, which substitutes the scalar calculation for the time-consuming calculation of vector similarity with several measures such as the squared distance, inner product, cosine, and minimum variance. Experimental results show that it makes clustering speed significantly fast for very high dimensional data.

• Proceedings of the KIEE Conference
• /
• 2008.07a
• /
• pp.303-304
• /
• 2008

For the state estimation problem, the weighted least squares (WLS) method and the fast decoupled method are widely used at present. However, these algorithms can converge to local optimal solutions. Recently, modern heuristic optimization methods such as Particle Swarm Optimization (PSO) have been introduced to overcome the disadvantage of the classical optimization problem. However, heuristic optimization methods based on populations require a lengthy computing time to find an optimal solution. In this paper, we used PSO to search for the optimal solution of state estimation in power systems. To overcome the shortcoming of heuristic optimization methods, we proposed parallel processing of the PSO algorithm based on the PC cluster system. the proposed approach was tested with the IEEE-118 bus systems. From the simulation results, we found that the parallel PSO based on the PC cluster system can be applicable for power system state estimation.

• Transactions of the Society of Information Storage Systems
• /
• v.5 no.2
• /
• pp.89-95
• /
• 2009

We setup the system to measure the noise characteristics of the 5M complementary metal-oxide semiconductor (CMOS) image sensor by generic measurement indicator of Standard mobile imaging architecture (SMIA) which is one of internal standard of mobile imaging architecture. To evaluate the effect of environment and setting parameters, such as temperature and integration time, we measure the variation of the dark signal, dynamic range and fixed pattern noise of image sensor. We also detect the number of defective pixels and cluster defects defined as adjacent single defect pixels at 5M CMOS image sensor. Then, we find the existence of some cluster defects in experiment, which are not expected in calculation.

• Proceedings of the Korea Society for Industrial Systems Conference
• /
• 2008.10b
• /
• pp.164-169
• /
• 2008

The increasing late-payment rate of credit card customers caused by a recent economic downturn are incurring not only reduced profit of department stores but also significant loss. Under this pressure, the objective of credit forecasting is extended from presumption of good or bad customers to contribution to revenue growth. As a method of managing defaults of department store credit card, this study classifies credit delinquents into some clusters, analyzes repaying patterns of customers in each cluster, and develops credit forecasting system to manage delinquents of department store credit card using data of Korean D department store's delinquents. The model presented by this study uses Kohonen network, a kind of artificial neural network of data mining techniques to cluster credit delinquents into groups. Logistic regression model is also used to predict repayment rate of customers of each cluster per period. The accuracy of presented system for the whole clusters is 92.3%.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 1996.10a
• /
• pp.290-293
• /
• 1996

This paper proposes a scheme of initial cluster center selection in FCM algorithm using the genetic algorithms. The FCM algorithm often fails in the search for global optimum because it is local search techniques that search for the optimum by using hill-climbing procedures. To solve this problem, we search for a hypersphere encircling each clusters whose parameters are estimated by the genetic algorithms. Then instead of a randomized initialization for fuzzy partition matrix in FCM algorithm, we initialize each cluster center by the center of a searched hypersphere. Our experimental results show that the proposed initializing scheme has higher probabilities of finding the global or near global optimal solutions than the traditional FCM algorithm.