• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2003.05a
• /
• pp.235-238
• /
• 2003

Fuzzy C-Means(FCM) 알고리즘은 초기 군집 중심의 개수와 위치에 따라 군집 결과의 성능차이가 많이 나타난다. 하지만 일반적인 경우에 군집 중심의 개수는 분석가의 주관에 의해 결정되고, 임의적으로 결정되기 때문에 원래 데이터의 구조와는 무관하게 수행되어 최적화된 군집화 수행을 실행하지 못하는 경우가 발생하게 된다. 따라서 본 논문에서는 원래의 데이터의 구조에 좀더 근접한 퍼지 군집화를 수행하기 위하여 격자를 바탕으로 한 데이터의 밀도를 이용한 FCM을 제안하고, 이러한 밀도 기반 FCM에 의해 결정된 군집의 합병 기법을 제안하였다. N-차원의 데이터 공간을 N-차원의 격자로 나누고, 초기 군집 중심의 개수와 위치는 각 격자의 밀도를 바탕으로 결정된다. 초기화 이후에 각 격자 내부에서 FCM을 이용하여 군집화를 수행하고, 계속해서 이웃 격자의 군집결과에 대하여 군집간의 유사도 측도를 이용하여 군집 합병을 수행함으로써 데이터의 자연적인 구조에 근접한 군집화를 수행하였다. 제안된 군집화 합병 기법의 향상된 성능은 UCI Machine Learning Repository 데이터를 이용하여 확인하였다.

• Journal of The Korean Society of Grassland and Forage Science
• /
• v.24 no.4
• /
• pp.293-300
• /
• 2004

This study was carried out to investigate the ecological behaviour forage value and similarity among the plant communities of the grazing pasture near Witzenhausen in middle part of Germany. Sixteen plant communities of the different grazing pasture were mostly the Molinio-Arrhenatheretea and Festuco-Brometea, and those were named the class of plant sociological nomenclature. The ecological behaviour and forage value of the communities except mesobromion(half dry grassland community) were relatively good for forage production. The correlation coefficient between class No. 14 and 12 of plant communities was highest, and the similarity among the communities were greatly affected by botanical composition. The resemblance measure of the cluster-analysis by complete-linkage-method for the similarity among plant communities was better the euclidean distance than those of others. The clustering analysis showed that the communities of relatively similar botanical composition were closely grouped.

• The KIPS Transactions:PartB
• /
• v.11B no.4
• /
• pp.485-490
• /
• 2004

In data representation, the clustering performs a grouping process which combines given data into some similar clusters. The various similarity measures have been used in many researches. But, the validity of clustering results is subjective and ambiguous, because of difficulty and shortage about objective criterion of clustering. The fuzzy clustering provides a good method for subjective clustering problems. It performs clustering through the similarity matrix which has fuzzy membership value for assigning each object. In this paper, for objective fuzzy clustering, the clustering algorithm which joins principal components analysis as a dimension reduction model with bayesian learning as a statistical learning theory. For performance evaluation of proposed algorithm, Iris and Glass identification data from UCI Machine Learning repository are used. The experimental results shows a happy outcome of proposed model.

• The KIPS Transactions:PartB
• /
• v.14B no.4
• /
• pp.263-272
• /
• 2007

This paper proposes a skeleton-based hierarchical shape description scheme, called a skeleton tree, for accurate shape-based image retrieval. A skeleton tree represents an object shape as a hierarchical tree where high-level nodes describe parts of coarse trunk regions and low-level nodes describe fine details of boundary regions. Each node refines the shape of its parent node. Most of the noise disturbances are limited to bottom level nodes and the boundary noise is reduced by decreasing weights on the bottom levels. The similarity of two skeleton trees is computed by considering the best match of a skeleton tree to a sub-tree of another skeleton tree. The proposed method uses a hybrid similarity measure by employing both Fourier descriptors and moment invariants in computing the similarity of two skeleton trees. Several experimental results are presented demonstrating the validity of the skeleton tree scheme for the shape description and indexing.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2011.05a
• /
• pp.93-96
• /
• 2011

본 논문에서는 개선된 FCM 알고리즘을 적용하여 통계청에서 제공하는 한국 표준 질병 사인 분류표(K.C.D)를 기초로 질병을 분류한 후, 질병을 도출하고 애매한 증상의 차이의 정도를 퍼지 추론기법을 사용하여 정확한 질병 상세를 도출할 수 있는 한방 질병 분류 시스템을 제시한다. 기존의 FCM 알고리즘은 입력 벡터들과 각 군집 중심과의 거리를 이용하여 측정된 유사도에 기초한 목적 함수의 최적화 방식을 사용한다. 하지만 측정된 패턴과 군집 공간상의 패턴들의 분포에 따라 바람직하지 못한 군집화 결과를 보일 수 있다. 따라서 본 논문에서는 군집들의 대칭성 측도에 퍼지 이론을 적용하여 기존의 FCM 알고리즘으로 군집화 한 결과를 재 군집화 하여 군집화의 정확성을 개선시킨 후, 증상의 차이를 구분하기 위해서 애매한 증상의 정도를 퍼지 추론 방법을 적용하여 정확한 질병 상세를 도출할 수 있는 방법을 제시한다. 본 논문에서는 개선된 FCM 알고리즘을 적용하여 질병을 분류한 후, 퍼지 제어 기법으로 질병을 추출함으로써 기존의 한방 자가진단 시스템 보다 정확하게 질병을 도출한 것을 확인하였다.

• The Journal of the Korea Contents Association
• /
• v.6 no.5
• /
• pp.20-28
• /
• 2006

This paper proposes visual-feature extraction for each band in wavelet domain with both spatial frequency features and multi resolution features, and the combination of visual features using fuzzy integral. In addition, it uses color feature expression method taking advantage of the frequency of the same color after color quantization for reducing quantization error, a disadvantage of the existing color histogram intersection method. Also, it is found that the final similarity can be represented in a linear combination of the respective factors(Homogram, color, energy) when each factor is independent one another. With respect to the combination patterns the fuzzy measurement is defined and the fuzzy integral is taken. Experiments are peformed on a database containing 1,000 color images. The proposed method gives better performance than the conventional method in both objective and subjective performance evaluation.

• The Journal of the Korea Contents Association
• /
• v.7 no.6
• /
• pp.22-30
• /
• 2007

Most real word systems such as world economy, stock market, and medical applications, contain a series of dynamic and complex phenomena. One of common methods to understand these systems is to build a model and analyze the behavior of the system. In this paper, we investigated methods for best clustering over time series data. As a first step for clustering, BIC (Bayesian Information Criterion) approximation is used to determine the number of clusters. A search technique to improve clustering efficiency is also suggested by analyzing the relationship between data size and BIC values. For clustering, two methods, model-based and similarity based methods, are analyzed and compared. A number of experiments have been performed to check its validity using real data(stock price). BIC approximation measure has been confirmed that it suggests best number of clusters through experiments provided that the number of data is relatively large. It is also confirmed that the model-based clustering produces more reliable clustering than similarity based ones.

• The Korean Journal of Applied Statistics
• /
• v.22 no.2
• /
• pp.415-424
• /
• 2009

As one of the unsupervised learning neural network methods, self-organizing maps(SOM) are applied to various fields. It reduces the dimension of multidimensional data by representing observations on the low dimensional manifold. On the other hand, the minimal spanning tree(MST) of a graph that achieves the most economic subset of edges connecting all components by a single open loop. In this study, we apply the MST technique to SOM with subnodes. We propose SOM's with embedded MST and a distance measure for optimum choice of the size and shape of the map. We demonstrate the method with Fisher's Iris data and a real gene expression data. Simulated data sets are also analyzed to check the validity of the proposed method.

• Journal of the Korean Data and Information Science Society
• /
• v.20 no.4
• /
• pp.615-627
• /
• 2009

In micro marketing promotion, it is important to know the behavior of customers. In this study we are interested in the forecasting of repurchase of customers from customers' behavior. By analyzing the cosmetic transaction data we derive some variables which play an important role in the knowledge of the customers' behavior and in the modeling of repurchase. As modeling tools we use the decision tree, logistic regression and neural network model. Finally we decide to use the decision tree as a final model since it yields the smallest RASE (root average squared error) and the greatest correct classification rate.

• Journal of the Korean Society for Marine Environment & Energy
• /
• v.3 no.2
• /
• pp.25-32
• /
• 2000

Fire-resistant boom is one of the most important facilities in in situ homing of spilled oil. Thermal response of a fire-resistant boom to turning is experimentally investigated in this paper by using an electric furnace and a burning test facility. This test facility is composed of a test tank, a fire boom, a hood for inhaling smoke, an incinerator for burning up gases and thermocouples, etc. Thereby a systematic method of approach in small laboratory scale is developed to study the performance of a fire-resistant boom. Burning test is carried out for the fire boom model which has been developed through the present study. It is shown that the present fire boom model has capability to withstand the high temperature around 800℃ and high rate of heat flux on it due to homing. For more realistic experimental environments, larger dimensions in devices and longer time in experiments are recommended in near future.