• Communications for Statistical Applications and Methods
• /
• 제6권3호
• /
• pp.781-790
• /
• 1999

In this paper we consider estimation of cancer incidence rates for local areas. The raw estimates usually are based on small sample sizes and hence are usually unreliable. A hierarchical Bayes generalized linear model is used which connects the local areas thereby enabling one to 'borrow strength' Random effects with pairwise difference priors model the spatial structure in the data. The methods are applied to cancer incidence estimation for census tracts in a certain region of the state of New York.

• Communications for Statistical Applications and Methods
• /
• 제17권3호
• /
• pp.309-318
• /
• 2010

The construction of asymptotic confidence intervals is considered for the difference of binomial proportions in two doubly sampled data subject to false-positive error. The coverage behaviors of several likelihood based confidence intervals and a Bayesian confidence interval are examined. It is shown that a hierarchical Bayesian approach gives a confidence interval with good frequentist properties. Confidence interval based on the Rao score is also shown to have good performance in terms of coverage probability. However, the Wald confidence interval covers true value less often than nominal level.

• 대한인간공학회:학술대회논문집
• /
• 대한인간공학회 1996년도 추계학술대회논문집
• /
• pp.181-184
• /
• 1996

Fuzzy hierarchical analysis(FHA) has the usefulness to allow decision maker's ambiguities when comparing two alternatives. But, for easiuly appling it to a decision problem, the handling its many data and for decision makers much not knowing fuzzy theory are the obstacles to must be overcomed even if the results of final fuzzy weights can be computed by a personal computer. This paper decribes that FHA is revised, and input/output interfaces are developed to collect input data easily and interprete the fuzzy resultlts. Finally, a fuzzy decision process is suggested with them.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제5권2호
• /
• pp.344-358
• /
• 2011

Wireless tactical network (WTN) is the most important present-day technology enabling modern network centric warfare. It inherits many features from WMNs, since the WTN is based on existing wireless mesh networks (WMNs). However, it also has distinctive characteristics, such as hierarchical structures and tight QoS (Quality-of-Service) requirements. Little research has been conducted on hierarchical protocols to support various QoS in WMN. We require new protocols specifically optimized for WTNs. Control packets are generally required to find paths and reserve resources for QoS requirements, so data throughput is not degraded due to overhead. The fundamental solution is to adopt topology aggregation, in which a low tier node aggregates and simplifies the topology information and delivers it to a high tier node. The overhead from control packet exchange can be reduced greatly due to decreased information size. Although topology aggregation is effective for low overhead, it also causes the inaccuracy of topology information; thus, incurring low QoS support capability. Therefore, we need a new topology aggregation algorithm to achieve high accuracy. In this paper, we propose a new aggregation algorithm based on star topology. Noting the hierarchical characteristics in military and hierarchical networks, star topology aggregation can be used effectively. Our algorithm uses a limited number of bypasses to increase the exactness of the star topology aggregation. It adjusts topology parameters whenever it adds a bypass. Consequently, the result is highly accurate and has low computational complexity.

• Management Science and Financial Engineering
• /
• 제12권1호
• /
• pp.65-77
• /
• 2006

This study proposes a Bayesian dynamic model in a hierarchical way to assess the time-varying effect of risk factors on the likelihood of corporate bankruptcy. For the longitudinal data, we aim to describe dynamically evolving effects of covariates more articulately compared to the Generalized Estimating Equation approach. In the analysis, it is shown that the proposed model outperforms in terms of sensitivity and specificity. Besides, the usefulness of this study can be found from the flexibility in describing the dependence structure among time specific parameters and suitability for assessing the time effect of risk factors.

• 한국멀티미디어학회논문지
• /
• 제23권1호
• /
• pp.1-7
• /
• 2020

Clustering methods essentially take a two-step approach; extracting feature vectors for dimensionality reduction and then employing clustering algorithm on the extracted feature vectors. However, for clustering images, the traditional clustering methods such as stacked auto-encoder based k-means are not effective since they tend to ignore the local information. In this paper, we propose a method first to effectively reduce data dimensionality using convolutional auto-encoder to capture and reflect the local information and then to accurately cluster similar data samples by using a hierarchical clustering approach. The experimental results confirm that the clustering results are improved by using the proposed model in terms of clustering accuracy and normalized mutual information.

• 한국정보시스템학회지:정보시스템연구
• /
• 제14권3호
• /
• pp.17-22
• /
• 2005

Location management is very important in location-based services to provide services to the mobile users like banking, city guides and many more. Ubiquitous and mobile devices are the source of data in location management and its significant operations are update and search method. Some studies to improve these were presented by using optimal sequential paging, location area scheme and hierarchical database scheme. In addition, not all location services have the same access methods on data and it lead to difficulties of providing services. A proposed location management of multi-agent architecture is presented in this study. It shows the coordination of the agents on the distributed database of location-based services. The proposal focuses on the location management of the mobile object presented in a hierarchical search and update. Also, it uses a nearest neighbor technique for efficient search method of mobile objects.

• 한국지능시스템학회논문지
• /
• 제13권5호
• /
• pp.512-519
• /
• 2003

본 논문에서는 계층적 클러스터링과 GMM을 순차적으로 이용하여 최적의 파라미터를 추정하고 이를 뉴로-퍼지 모델의 초기 파리미터로 사용하여 모델의 성능 개선을 제안한다. 반복적인 시도 중 가장 좋은 파라미터를 선택하는 기존의 알고리즘 과 달리 계층적 클러스터링은 데이터들 간의 유클리디언 거리를 이용하여 클러스터를 생성하므로 반복적인 시도가 불필요하다. 또한 클러스터링 방법에 의해 퍼지 모델링을 행하므로 클러스터와 동일한 갯수의 적은 규칙을 갖는다. 제안된 방법의 유용함을 비선형 데이터인 Box-Jenkins의 가스로 예측 문제와 Sugeno의 비선형 시스템에 적용하여 이전의 연구보다 적은 규칙으로도 성능이 개선되는 것을 보였다.

• Journal of Information Processing Systems
• /
• 제16권3호
• /
• pp.733-741
• /
• 2020

A smart dust monitoring system is useful for obtaining information on rough terrain that is difficult for humans to access. One of ways to deploy sensors to gather information in smart dust environment is to use an aircraft in the Amazon rainforest to scatter an enormous amount of small and cheap sensors (or smart dust devices), or to use an unmanned spacecraft to throw the sensors on the moon's surface. However, scattering an enormous amount of smart dust devices creates the difficulty of managing such devices as they can be scattered into inaccessible areas, and also causes problems such as bottlenecks, device failure, and high/low density of devices. Of the various problems that may occur in the smart dust environment, this paper is focused on solving the bottleneck problem. To address this, we propose and construct a three-layered hierarchical smart dust monitoring system that includes relay dust devices (RDDs). An RDD is a smart dust device with relatively higher computing/communicating power than a normal smart dust device. RDDs play a crucial role in reducing traffic load for the system. To validate the proposed system, we use climate data obtained from authorized portals to compare the system with other systems (i.e., non-hierarchical system and simple hierarchical system). Through this comparison, we determined that the transmission processing time is reduced by 49%-50% compared to other systems, and the maximum number of connectable devices can be increased by 16-32 times without compromising the system's operations.

• 정보보호학회논문지
• /
• 제18권1호
• /
• pp.49-57
• /
• 2008

본 논문에서는 구조적으로 분산 침입탐지시스템의 구조를 계승하면서 동시에 홈네트워크의 환경을 최대한 고려하여 HNHDIDS(Home Network Hierarchical Distributed Intrusion Detection System)로 명명된 새로운 계층적 분산 침입탐지 시스템을 제안한다. 제안된 시스템은 단일 클래스 support vector machine(support vector data description)과 지역적 에이전트(agent)들을 계층적으로 결합한 구조로써, 홈네트워크의 환경을 위하여 최적화되었다. 만족스러운 침입 탐지율과 안전한 FNR(false negative rate) 수치 등을 실험을 통하여 확인함으로써 제안된 시스템이 홈네트워크 환경에 적합함을 검증하였다.