• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.32 no.11A
• /
• pp.1164-1172
• /
• 2007

To solve scalability problem in the ZigBee Network, this paper presents a new mesh routing protocol for ZigBee, called ZigBee Cluster Label (ZiCL). ZiCL divides the ZigBee network into one or more logical clusters and then assigns a unique Cluster Label to each cluster so that it discovers a route of a destination node based on Cluster Label. When a node collects new Cluster Label information of a destination node according to discovery based on Cluster Label, ZiCL encourages nodes with the same Cluster Label to share the information. Consequen tly, it contributes on reducing numerical potential route discoveries and improving network performances such as routing overhead, end-to-end delay, and packet delivery ratio. Simulation results using NS-2 show ZiCL performs well.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.33 no.8B
• /
• pp.689-698
• /
• 2008

The performance of a TCP protocol on MANETs has been studied in a numerous researches. One of the significant reasons of TCP performance degradation on MANETs is inability to distinguish between packet losses due to congestion from those caused by nodes mobility and as consequence broken routes. This paper presents the Cluster-Label-based Routing (CLR) protocol that is an attempt to compensate source of TCP problems on MANETs - multi-hop mobile environment. By utilizing Cluster-Label-based mechanism for Backbone, the CLR is able to concentrate on detection and compensation of movement of a destination node. The proposed protocol provides better goodput and delay performance than standardized protocols especially in cases of large network size and/or high mobility rate.

• Journal of Institute of Control, Robotics and Systems
• /
• v.18 no.1
• /
• pp.35-41
• /
• 2012

This paper proposes a moving object detection algorithm for active camera system that can be applied to mobile robot and intelligent surveillance system. Most of moving object detection algorithms based on a stationary camera system. These algorithms used fixed surveillance system that does not consider the motion of the background or robot tracking system that track pre-learned object. Unlike the stationary camera system, the active camera system has a problem that is difficult to extract the moving object due to the error occurred by the movement of camera. In order to overcome this problem, the motion of the camera was compensated by using SURF and Pseudo Perspective model, and then the moving object is extracted efficiently using stochastic Label Cluster transport model. This method is possible to detect moving object because that minimizes effect of the background movement. Our approach proves robust and effective in terms of moving object detection in active camera system.

• Journal of Internet Computing and Services
• /
• v.18 no.2
• /
• pp.61-73
• /
• 2017

In this paper, we propose a Documents Cluster Labeling method using information content of words in clusters to understand what the clusters imply. To do so, we calculate the weight and frequency of the words. These two measures are used to determine the weight among the words in the cluster. As a nest step, we identify the candidate labels using the WordNet. At this time, the candidate labels are matched to least common hypernym of the words in the cluster. Finally, the representative labels are determined with respect to information content of the words and the weight of the words. To prove the superiority of our method, we perform the heuristic experiment using two kinds of measures, named the suitability of the candidate label ($Suitability_{cl}$) and the appropriacy of representative label ($Appropriacy_{rl}$). In applying the method proposed in this research, in case of suitability of the candidate label, it decreases slightly compared with existing methods, but the computational cost is about 20% of the conventional methods. And we confirmed that appropriacy of the representative label is better results than the existing methods. As a result, it is expected to help data analysts to interpret the document cluster easier.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.33 no.11A
• /
• pp.1034-1042
• /
• 2008

It is very hard, but important to sustain path stability for a reliable communication in mobile ad hoc networks. We propose a novel source routing protocol that establishes a group path with virtual multiple paths to enable a robust communication. The entire mobile nodes form a disjoint set of clusters: Each has its clusterhead as a cluster leader and a unique cluster label to identify itself from other clusters. A group path is a sequence of cluster labels instead of nodes and the nodes with the same label collaborate to deliver packets to a node with next label on the group path. We prove by resorting to simulation that our proposed protocol outperforms the existing key routing protocols, even for a network with a high node mobility and a high traffic.

• Agribusiness and Information Management
• /
• v.7 no.1
• /
• pp.21-36
• /
• 2015

This study aims to analyze the differences of ginseng product consumers and segment Korean fresh ginseng and red ginseng root markets based on attributes for the purchase. As a result of analyzing survey data, the red ginseng root consumers had different aspects from fresh ginseng consumers. According to the result of cluster analysis, the fresh ginseng consumers were subdivided into three segments (safety-oriented consumption cluster, label centered consumption cluster, and high involvement consumption cluster), while the red ginseng root consumers were subdivided into four segments (convenience-oriented consumption cluster, high involvement consumption cluster, raw material's safety-oriented cluster, and raw material's information importance cluster). ANOVA and Crosstab were conducted to investigate characteristics of each cluster.

• 한국어정보학회:학술대회논문집
• /
• 2017.10a
• /
• pp.56-59
• /
• 2017

CNN(Convolutional Neural Network)을 이용하여 발화 주제 다중 분류 task를 multi-labeling 방법과, cluster 방법을 이용하여 수행하고, 각 방법론에 MSE(Mean Square Error), softmax cross-entropy, sigmoid cross-entropy를 적용하여 성능을 평가하였다. Network는 음절 단위로 tokenize하고, 품사정보를 각 token의 추가한 sequence와, Naver DB를 통하여 얻은 named entity 정보를 입력으로 사용한다. 실험결과 cluster 방법으로 문제를 변형하고, sigmoid를 output layer의 activation function으로 사용하고 cross entropy cost function을 이용하여 network를 학습시켰을 때 F1 0.9873으로 가장 좋은 성능을 보였다.

• Journal of the Korea Society of Computer and Information
• /
• v.26 no.1
• /
• pp.27-35
• /
• 2021

In this paper, we propose a new stacking ensemble framework for deep learning models which reflects the distribution of label embeddings. Our ensemble framework consists of two phases: training the baseline deep learning classifier, and training the sub-classifiers based on the clustering results of label embeddings. Our framework aims to divide a multi-class classification problem into small sub-problems based on the clustering results. The clustering is conducted on the label embeddings obtained from the weight of the last layer of the baseline classifier. After clustering, sub-classifiers are constructed to classify the sub-classes in each cluster. From the experimental results, we found that the label embeddings well reflect the relationships between classification labels, and our ensemble framework can improve the classification performance on a CIFAR 100 dataset.

• Journal of information and communication convergence engineering
• /
• v.11 no.3
• /
• pp.179-184
• /
• 2013

Traditional clustering methods are usually based on the bag-of-words (BOW) model. A disadvantage of the BOW model is that it ignores the semantic relationship among terms in the data set. To resolve this problem, ontology or matrix factorization approaches are usually used. However, a major problem of the ontology approach is that it is usually difficult to find a comprehensive ontology that can cover all the concepts mentioned in a collection. This paper proposes a new document clustering method using semantic features and fuzzy relations for solving the problems of ontology and matrix factorization approaches. The proposed method can improve the quality of document clustering because the clustered documents use fuzzy relation values between semantic features and terms to distinguish clearly among dissimilar documents in clusters. The selected cluster label terms can represent the inherent structure of a document set better by using semantic features based on non-negative matrix factorization, which is used in document clustering. The experimental results demonstrate that the proposed method achieves better performance than other document clustering methods.

• Bulletin of the Korean Chemical Society
• /
• v.34 no.3
• /
• pp.815-819
• /
• 2013

Micropatterns of streptavidin and human epidermal carcinoma A431 cells were successfully imaged, as received and without any labeling, using cluster $Au_3{^+}$ ion beam-based time-of-flight secondary ion mass spectrometry (TOF-SIMS) together with a principal component analysis (PCA). Three different analysis ion beams ($Ga^+$, $Au^+$ and $Au_3{^+}$) were compared to obtain label-free TOF-SIMS chemical images of micropatterns of streptavidin, which were subsequently used for generating cell patterns. The image of the total positive ions obtained by the $Au_3{^+}$ primary ion beam corresponded to the actual image of micropatterns of streptavidin, whereas the total positive-ion images by $Ga^+$ or $Au^+$ primary ion beams did not. A PCA of the TOF-SIMS spectra was initially performed to identify characteristic secondary ions of streptavidin. Chemical images of each characteristic ion were reconstructed from the raw data and used in the second PCA run, which resulted in a contrasted - and corrected - image of the micropatterns of streptavidin by the $Ga^+$ and $Au^+$ ion beams. The findings herein suggest that using cluster-ion analysis beams and multivariate data analysis for TOF-SIMS chemical imaging would be an effectual method for producing label-free chemical images of micropatterns of biomolecules, including proteins and cells.