• Journal of KIISE:Computing Practices and Letters
• /
• v.16 no.4
• /
• pp.482-486
• /
• 2010

Recently, gesture recognition methods based on dynamic time warping (DTW) have been actively investigated as more mobile devices have equipped the accelerometer. DTW has no additional training step since it uses given samples as the matching templates. However, it is difficult to apply the DTW on mobile environments because of its computational complexity of matching step where the input pattern has to be compared with every templates. In order to address the problem, this paper proposes a gesture recognition method based on DTW that uses localized subset of templates. Here, the k-means clustering algorithm is used to divide each class into subclasses in which the most centered sample in each subclass is employed as the localized template. It increases the recognition speed by reducing the number of matches while it minimizes the errors by preserving the diversities of the training patterns. Experimental results showed that the proposed method was about five times faster than the DTW with all training samples, and more stable than the randomly selected templates.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.30 no.5B
• /
• pp.295-303
• /
• 2005

Recently, according to the rapid increase of Web users, various kinds of Web applications have been being developed. Hence, Web QoS(Quality of Service) becomes a critical issue in the Web services, such as e-commerce, Web hosting, etc. Nevertheless, most Web servers currently process various requests from Web users on a FIFO basis, which can not provide differentiated QoS. This paper presents a load balancing method to provide differentiated Web QoS in clustering web server. The first is the kernel-level approach, which is adding a real-time scheduling process to the operating system kernel to maintain the priority of user requests determined by the scheduling process of Web server. The second is the load-balancing approach, which uses IP-level masquerading and tunneling technology to improve reliability and response speed upon user requests. The third is the dynamic load-balancing approach, which uses the parameters related to the MIB-II of SNMP and the parameters related to load of the system such as memory and CPU.

• Journal of Internet Computing and Services
• /
• v.17 no.6
• /
• pp.103-111
• /
• 2016

Offloading tasks from mobile devices to available cloud servers were improved since the introduction of the cloudlet. With the implementation of dynamic offloading algorithms, mobile devices can choose the appropriate server for the set of tasks. However, current task distribution approaches do not consider the number of VM, which can be a critical factor in the decision making. This paper proposes a dynamic task distribution on clustered data centers. A proportional VM migration approach is also proposed, where it migrates virtual machines to the cloud servers proportionally according to their allocated CPU, in order to prevent overloading of resources in servers. Moreover, we included the resource capacity of each data center in terms of the maximum CPU in order to improve the migration approach in cloud servers. Simulation results show that the proposed mechanism for task distribution greatly improves the overall performance of the system.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2003.11a
• /
• pp.165-168
• /
• 2003

네트워크 시스템이 발달하면서 다양한 컴퓨터들을 연결하는 클러스터링 시스템 구축이 용이해졌다. 이러한 이기종 클러스터 환경을 구축함에 있어서 노드들간의 성능 분균형으로 인한 문제가 야기되는데 본 논문에서는 Message Passing 방식을 이용한 클러스터링을 구축함에 있어서 노드들의 자원의 정보를 이용하여 메모리의 과부하를 최대한 예방하여 작업을 메모리 여유가 있는 노드로 이주시킴으로써 시스템 안정성과 자원을 균등하게 사용할 수 있도록 제안하였다. 제안한 알고리즘을 구현하기 위해서 이기종 클러스터 환경에서 MPI를 이용하여 2차원 열에너지 전도 계산과 Matrix 곱셈 프로그램을 이용하여 제안한 알고리즘과 GSS, Send 알고리즘, Weighted Factoring알고리즘들과 상대 비교를 하였다.

• Proceedings of the Korean Information Science Society Conference
• /
• 1998.10c
• /
• pp.147-149
• /
• 1998

본 논문에서는 여러 정보검색 엔진들이 분산되어 있는 환경에서 이 엔진들의 검색 결과를 효과적으로 취합하여 사용자에게 제시하는 컬렉션 융합 방안을 제안하고자 한다. 이 방법은 우선 학습 질의어로 검색된 문서들의 클러스터링 정도를 이용하여 컬렉션에의 신뢰도를 측정하고 새로운 질의어가 입력되었을 때 각 컬렉션에서 검색된 문서의 유사도를 조정하여 융합하는 방법이다. 여기에서 각 컬렉션의 신뢰도는 미리 준비된 학습 질의어와 이 학습 질의어를 입력하여 검색된 문서들 사이의 유사도를 분석하여 측정한다. 이 신뢰도는 새로운 질의어가 입력되었을 때 각 컬렉션마다 문서들을 검색하고 이들 문서들을 어느 정도 신뢰할 것인가를 결정하는데 사용된다. 본 논문에서 제안한 방법은 학습과정에서 사람이 학습시킬 필요가 없는 비지도 학습에 기초하고 있다. 따라서 지금까지 지도 학습에 기초한 컬렉션 융합 방법과는 달리 인터넷과 같이 문서들이 동적으로 변하는 환경에서 쉽게 사용할 수 있다는 장점을 가진다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2005.07a
• /
• pp.28-30
• /
• 2005

센서들은 제한된 자원으로 구동되므로 오류가 나기 쉽다. 특히 구조적 라우팅의 경우 클러스터 헤드의 오류시 많은 수의 센서가 네트워크에서 분리되어 네트워크 성능에 악영향을 미친다. 따라서 오류 처리에 관한 연구들이 이루어져 왔으나 기존의 연구들은 망을 최적으로 유지하기 위해 재구성시 전체 네트워크를 재구성 하며 고정된 주기를 사용하여 전체적인 망이 최적의 상태임에도 불구하고 재구성 되거나, 클러스터에 오류가 생겨도 재구성되기까지 기다려야 한다는 단점이 있었다. 따라서 본 논문에서는 지역적인 재클러스터링을 통하여 네트워크를 최적으로 유지하며 클러스터들의 부하를 고려하여 망을 동적으로 재구성 하는 방법을 제안하였다. NS-2를 이용한 시뮬레이션을 통하여 기존의 방법에 비하여 본 논문에서 제안한 알고리즘이 네트워크 유지시간을 연장시켜 더 많은 양의 데이터가 수집됨을 확인 할 수 있었다.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2009.05a
• /
• pp.576-579
• /
• 2009

The amount of incoming e-mails is increasing rapidly due to the wide usage of Internet. Therefore, it is more required to classify incoming e-mails efficiently and accurately. Currently, the e-mail classification techniques are focused on two way classification to filter spam mails from normal ones based mainly on Bayesian and Rule. The clustering method has been used for the multi-way classification of e-mails. But it has a disadvantage of low accuracy of classification. In this paper, we propose a novel multi-way e-mail classification method that uses PCA for automatic category generation and dynamic category hierarchy for high accuracy of classification. It classifies a huge amount of incoming e-mails automatically, efficiently, and accurately.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.17 no.2
• /
• pp.495-502
• /
• 2013

In the mobile Ad hoc Network(MANET), improving technique for management and control of topology is recognized as an important part of the next generation network. In this paper, we proposed an efficient node life time management of ATICC(Adaptive Time Interval Clustering Control) in Ad-hoc Networks. Ad-hoc Network is a self-configuration network or wireless multi-hop network based on inference topology. This is a method of path routing management node for increasing the network life time through the periodical route alternation. The proposed ATICC algorithm is time interval control technique depended on the use of the battery energy while node management considering the attribute of node and network routing. This can reduce the network traffic of nodes consume energy cost effectively. As a result, it could be improving the network life time by using timing control method in ad-hoc networks.

• KIPS Transactions on Computer and Communication Systems
• /
• v.2 no.5
• /
• pp.199-202
• /
• 2013

In the cluster mechanism of the wireless sensor network, the network lifetime is affected by how cluster heads are selected. One of the representative clustering mechanisms, the low-energy adaptive clustering hierarchy (LEACH), selects cluster heads periodically, resulting in high energy consumption in cluster reconstruction. On the other hand, the adaptive clustering algorithm via waiting timer (ACAWT) proposes a non-periodic re-clustering mechanism that reconstructs clusters if the remaining energy level of a cluster head reaches a given threshold. In this paper, we propose a re-clustering mechanism that uses multiple remaining node energy levels and does re-clustering when the remaining energy level of a cluster head reaches one level lower. Also, in determining cluster heads, both of the number of neighbor nodes and the remaining energy level are considered so that cluster heads can be more evenly placed. From the simulations based on the Qualnet simulator, we validate that our proposed mechanism outperforms ACAWT in terms of the network lifetime.

• Journal of KIISE
• /
• v.45 no.3
• /
• pp.288-293
• /
• 2018

There have recently been several studies investigating how to apply a privacy preserving technique to publish data. Differential privacy can protect personal information regardless of an attacker's background knowledge by adding probabilistic noise to the original data. To perform differentially private k-means clustering, the existing algorithm builds a differentially private histogram and performs the k-means clustering. Since it constructs an equi-width histogram without considering the distribution of data, there are many buckets to which noise should be added. We propose a k-means clustering algorithm using a quad-tree that captures the distribution of data by using a small number of buckets. Our experiments show that the proposed algorithm shows better performance than the existing algorithm.