• 한국경영과학회:학술대회논문집
• /
• 한국경영과학회/대한산업공학회 2003년도 춘계공동학술대회
• /
• pp.729-733
• /
• 2003

The presence of local features within clusters incurred by multi-modal nature of data prohibits many conventional clustering techniques from working properly. Especially, the clustering of datasets with non-Gaussian distributions within a cluster can be problematic when the technique with implicit assumption of Gaussian distribution is used. Current study proposes a simple tandem clustering method composed of k-means type algorithm and hierarchical method to solve such problems. The multi-modal dataset is first divided into many small pre-clusters by k-means or fuzzy k-means algorithm. The pre-clusters found from the first step are to be clustered again using agglomerative hierarchical clustering method with Kullback- Leibler divergence as the measure of dissimilarity. This method is not only effective at extracting the multi-modal clusters but also fast and easy in terms of computation complexity and relatively robust at the presence of outliers. The performance of the proposed method was evaluated on three generated datasets and six sets of publicly known real world data.

• 응용통계연구
• /
• 제22권4호
• /
• pp.745-758
• /
• 2009

자발적인 군집을 유도하는 다변량 통계기법으로 널리 사용되는 군집분석은 데이터에 기반한 탐색적 방법으로 쓰이며 군집원칙에 따라 여러 가지 방법이 제안되어 왔다. 또한 군집화된 결과에 대하여 유효성을 측정하는 측도도 다양한방법이 개발되었다. 본 연구에서는 계층적 군집분석 방법으로 최장연결법과 Ward의 방법, 비계층적 군집분석 방법으로 K-평균법 그리고 확률분포정보를 활용한 모형기반 군집분석방법을 이용하여 모의실험으로 군집분석을 실시하고 군집유효성 측도로는 연결성, Dunn 지수, 실루엣을 구하여 각 군집방법에 대해 유효성을 비교한다. 또한, 한우 관능평가 데이터에 군집분석을 적용하여 최적의 군집 상황을 구하고자 한다.

• 기술경영경제학회:학술대회논문집
• /
• 기술경영경제학회 2005년도 27회 하계학술발표회 논문집
• /
• pp.116-139
• /
• 2005

ICT(information and communication technology) has played a pivotal role in the world economy, and the out look for ICT has improved markedly. One of the noticeable characteristics in the ICT sector Is the global rationalization of its technology and service. Specialization on the specific ICT capability is a pressing problem for many countries. Along the line of classical innovation cluster and network studies, this paper suggests a way to find and analyze the role of core technologies on the ICT network First, technology network is constructed by using patent citation data from USPTO. Then, a couple of cluster is generated by K-means clustering technique. Finally, brokerage analysis is applied to manifest the role of principal technologies. The network visualization and some stylized facts on dynamics are briefly given altogether Based on the role and relationship of technologies across clusters, it is expected that this research could contribute to the ICT cluster formation and the vision-making for ICT specialization at the viewpoint of technology Policy.

• 한국군사과학기술학회지
• /
• 제13권3호
• /
• pp.453-460
• /
• 2010

For high quality image acquisition, compensating air-vehicle motion is essential for airborne SAR system. This paper describes a timing control based motion compensation method for airborne SAR system. Efficient timing control is critical for SAR system since it maintains many timing signals and timing setting for the signals should be updated frequently. This paper proposes Timing Cluster method as an efficient means for timing control of SAR system. Moreover, this paper suggests a simple and efficient method to compensate air-vehicle motion based on the Timing Cluster method. Timing Cluster method enables SAR system to control the timing in a timing noncritical way just maintaining little amount of information.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1994년도 Proceedings of the Korea Automatic Control Conference, 9th (KACC) ; Taejeon, Korea; 17-20 Oct. 1994
• /
• pp.113-116
• /
• 1994

In this paper. we suggest new fuzzy modeling algorithm, which can be easily implemented, by combining HPC-MEANS Algorithm and Genetic Algorithm. HPC-MEANS used to cluster the sample data in input-output space will hyper planes and to make structure identification roughly and Genetic Algorithm is used to nine the premise and consequent parameters. For the validity of suggested methods we model the system with I/O data from known system. and then compare two systems.

• 센서학회지
• /
• 제13권6호
• /
• pp.446-453
• /
• 2004

A portable electronic tongue (E-Tongue) system using an array of ion-selective electrode (ISE) and personal digital assistants (PDA) for recognizing and analyzing food and drink have been designed. By the employment of PDA, the complex algorithm such as fuzzy c-means algorithm (FCMA) could be used in E-Tongue, PUMA could iteratively solve the cluster centers of pre-determined standard patterns. And the membership between the standard patterns and unknown pattern could be analyzed easily by the present E-Tongue combined with PDA.

• 한국전기전자재료학회논문지
• /
• 제27권12호
• /
• pp.877-882
• /
• 2014

It has been confirmed that the inner defect of transformer and the perfect diagnosis for aging are closely related to safe electric power transmission system and that the detection of accident and diagnosis technique turn out to be very important issues. Since electric power machinery consists of various kinds of components, however, it is very difficult to make a diagnosis for aging by one parameter. Thus, diagnosis for aging is feasible only through the combination of various parameters. Recently, various expert systems have been developed and applied to diagnosis for aging, but they are not yet reliable enough to apply to the real system. In this paper, XLPE which is ultra high voltage cable insulator material were chosen to investigate the influence of void on insulator material using partial discharge. Obtained data have been processed by PRPD (phased resolved partial discharge) distribution function and K-means. And statistical and cluster distribution of partial discharge have been analysed and investigated.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제4권6호
• /
• pp.1023-1041
• /
• 2010

In cognitive radio networks, cooperative spectrum sensing schemes are proposed to improve the performance of detecting licensees by secondary users. Commonly, the cooperative sensing can be realized by means of hard decision fusion (HDF) or soft decision fusion (SDF) schemes. The SDF schemes are superior to the HDF ones in terms of the detection performance whereas the HDF schemes are outperforming the SDF ones when the traffic overhead is taken into account. In this paper, a hybrid SFD-HDF cluster-based approach is developed to jointly exploit the advantages of SFD and HDF schemes. Different SDF schemes have been proposed and compared within a given cluster whereas the OR-rule base HDF scheme is applied to combine the decisions reported by cluster headers to a common receiver or base station. The computer simulations show promising results as the performance of the proposed scenario of hybridizing soft and hard fusion schemes is significantly outperforming other different combinations of conventional SDF and HDF schemes while it noticeably reduces the network traffic overhead.

• 한국음향학회지
• /
• 제26권5호
• /
• pp.199-206
• /
• 2007

본 논문에서는 화자식별을 위한 특징벡터의 새로운 클러스터링 방법을 제안한다. 제안된 방법은 클러스터 센터에 대한 초기값 설정과 클러스터 개수에 대한 사전 정보 없이 클러스터링이 가능하다. 각 클러스터 센터는 경계 차감 클러스터링 알고리즘으로 한 번에 한 개의 클러스터 센터가 추가됨으로써 순차적으로 구해지며, 클러스터 개수는 클러스터간의 상호관계를 조사하여 결정된다. 인공 생성 데이터 및 TIMIT 음성을 이용하여 실험한 결과로부터 제안된 방법이 기존의 방법보다 우수함을 확인하였다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제14권4호
• /
• pp.1773-1795
• /
• 2020

Unmanned aerial vehicles (UAVs) network are a very vibrant research area nowadays. They have many military and civil applications. Limited bandwidth, the high mobility and secure communication of micro UAVs represent their three main problems. In this paper, we try to address these problems by means of secure clustering, and a security clustering algorithm based on integrated trust value for UAVs network is proposed. First, an improved the k-means++ algorithm is presented to determine the optimal number of clusters by the network bandwidth parameter, which ensures the optimal use of network bandwidth. Second, we considered variables representing the link expiration time to improve node clustering, and used the integrated trust value to rapidly detect malicious nodes and establish a head list. Node clustering reduce impact of high mobility and head list enhance the security of clustering algorithm. Finally, combined the remaining energy ratio, relative mobility, and the relative degrees of the nodes to select the best cluster head. The results of a simulation showed that the proposed clustering algorithm incurred a smaller computational load and higher network security.