• Journal of the Korean Operations Research and Management Science Society
• /
• v.29 no.1
• /
• pp.43-56
• /
• 2004

The purpose of clustering is to partition a set of objects into several clusters based on some appropriate similarity measure. In most cases, clustering is considered without any prior information on the number of clusters or the structure of the given data, which makes clustering is one example of very complicated combinatorial optimization problems. In this paper we propose a general-purpose clustering method that can determine the proper number of clusters as well as efficiently carry out clustering analysis for various types of data. The method is composed of two stages. In the first stage, two different hierarchical clustering methods are used to get a reasonably good clustering result, which is improved In the second stage by ASA(accelerated simulated annealing) algorithm equipped with specially designed perturbation schemes. Extensive experimental results are given to demonstrate the apparent usefulness of our ASA clustering method.

• The Journal of the Acoustical Society of Korea
• /
• v.32 no.2
• /
• pp.174-180
• /
• 2013

This paper proposes a decision tree based context clustering algorithm for HMM-based speech synthesis systems using the cross likelihood ratio with a hierarchical prior (CLRHP). Conventional algorithms tie the context-dependent HMM states that have similar statistical characteristics, but they do not consider the statistical similarity of split child nodes, which does not guarantee the statistical difference between the final leaf nodes. The proposed CLRHP algorithm improves the reliability of model parameters by taking a criterion of minimizing the statistical similarity of split child nodes. Experimental results verify the superiority of the proposed approach to conventional ones.

• 한국데이터정보과학회:학술대회논문집
• /
• 2005.04a
• /
• pp.185-192
• /
• 2005

There are many data mining techniques such as association rule, decision tree, neural network analysis, clustering, genetic algorithm, bayesian network, memory-based reasoning, etc. We analyze 2003 Gyeongnam social indicator survey data using k-means clustering technique for environmental information. Clustering is the process of grouping the data into clusters so that objects within a cluster have high similarity in comparison to one another. In this paper, we used k-means clustering of several clustering techniques. The k-means clustering is classified as a partitional clustering method. We can apply k-means clustering outputs to environmental preservation and environmental improvement.

• Journal of Information Processing Systems
• /
• v.15 no.5
• /
• pp.1055-1067
• /
• 2019

In view of the deficiencies of existing weighted similarity indexes, a hierarchical clustering method initialize-expand-merge (IEM) is proposed based on the similarity of common neighbors for community discovery in weighted networks. Firstly, the similarity of the node pair is defined based on the attributes of their common neighbors. Secondly, the most closely related nodes are fast clustered according to their similarity to form initial communities and expand the communities. Finally, communities are merged through maximizing the modularity so as to optimize division results. Experiments are carried out on many weighted networks, which have verified the effectiveness of the proposed algorithm. And results show that IEM is superior to weighted common neighbor (CN), weighted Adamic-Adar (AA) and weighted resources allocation (RA) when using the weighted modularity as evaluation index. Moreover, the proposed algorithm can achieve more reasonable community division for weighted networks compared with cluster-recluster-merge-algorithm (CRMA) algorithm.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.9 no.1
• /
• pp.190-207
• /
• 2015

In the combustion engineering, to decrease pollution and increase production efficiency, and to optimally keep solid burning material amount constant in a burner online, it needs a smart method to detect the amount variation of the burning materials in a high temperature environment. This paper presents an online machine vision system for automatically measuring and detecting the burning material amount inside a burner or a boiler. In the camera-protecting box of the system, a sub-system for cooling is constructed by using the cooling water circulation techqique. In addition, the key and intelligent step in the system is to detect the pile profile of the variable burning material, and the algorithm for the pile profile tracing was studied based on the combination of the gey level (color) discontinuity and similarity based image segmentation methods, the discontinuity based sub-algorithm is made on the quaternion convolution, and the similarity based sub-algorithm is designed according to the region growing with multi-scale clustering. The results of the two sub-algoritms are fused to delineate the final pile profile, and the algorithm has been tested and applied in different industrial burners and boilers. The experiements show that the proposed algorithm works satisfactorily.

• Language and Information
• /
• v.14 no.1
• /
• pp.113-144
• /
• 2010

The primary goal of this paper is to find a feasible way to answer the question: Does the similarity in meaning between verbs relate to the similarity in their subcategorization? In order to answer this question in a rather concrete way on the basis of a large set of English verbs, this study made use of various language resources, tools, and statistical methodologies. We first compiled a list of 678 verbs that were selected from the most and second most frequent word lists from the Colins Cobuild English Dictionary, which also appeared in WordNet 3.0. We calculated similarity measures between all the pairs of the words based on the 'jcn' algorithm (Jiang and Conrath, 1997) implemented in the WordNet::Similarity module (Pedersen, Patwardhan, and Michelizzi, 2004). The clustering process followed, first building similarity matrices out of the similarity measure values, next drawing dendrograms on the basis of the matricies, then finally getting 177 meaningful clusters (covering 437 verbs) that passed a certain level set by z-score. The subcategorization frames and their frequency values were taken from the ICE-GB. In order to calculate the Selectional Preference Strength (SPS) of the relationship between a verb and its subcategorizations, we relied on the Kullback-Leibler Divergence model (Resnik, 1996). The SPS values of the verbs in the same cluster were compared with each other, which served to give the statistical values that indicate how much the SPS values overlap between the subcategorization frames of the verbs. Our final analysis shows that the degree of overlap, or the relationship between semantic similarity and the subcategorization frames of the verbs in English, is equally spread out from the 'very strongly related' to the 'very weakly related'. Some semantically similar verbs share a lot in terms of their subcategorization frames, and some others indicate an average degree of strength in the relationship, while the others, though still semantically similar, tend to share little in their subcategorization frames.

• Journal of Institute of Control, Robotics and Systems
• /
• v.16 no.8
• /
• pp.744-750
• /
• 2010

To segment instances of vehicle classes in a sparse stereo-matching data set, this paper presents an algorithm for clustering based on DP (Dynamic Programming). The algorithm is agglomerative: it begins with each element in the set as a separate cluster and merges them into successively larger clusters according to similarity of two clusters. Here, similarity is formulated as a cost function of DP. The proposed algorithm is proven to be effective by experiments performed on various images acquired by a moving vehicle.

• Journal of the Korean Society for information Management
• /
• v.21 no.1
• /
• pp.93-117
• /
• 2004

This study is to develop a hierarchic clustering model fur document classification and browsing in OPAC systems. Two automatic indexing techniques (with and without controlled terms), two term weighting methods (based on term frequency and binary weight), five similarity coefficients (Dice, Jaccard, Pearson, Cosine, and Squared Euclidean). and three hierarchic clustering algorithms (Between Average Linkage, Within Average Linkage, and Complete Linkage method) were tested on the document collection of 175 books and theses on library and information science. The best document clusters resulted from the Between Average Linkage or Complete Linkage method with Jaccard or Dice coefficient on the automatic indexing with controlled terms in binary vector. The clusters from Between Average Linkage with Jaccard has more likely decimal classification structure.

• Journal of Korea Multimedia Society
• /
• v.18 no.2
• /
• pp.138-148
• /
• 2015

Object detection methods based on background learning are widely used in video surveillance. However, when a car runs with headlights on, these methods are likely to detect the car region and the area illuminated by the headlights as one connected change region. This paper describes a method of separating the car region from the area illuminated by the headlights. First, we detect change regions with a background learning method, and extract blobs, connected components in the detected change region. If a blob is larger than the maximum object size, we extract candidate object regions from the blob by clustering the intensity histogram of the frame difference between the mean of background images and an input image. Finally, we compute the similarity between the mean of background images and the input image within each candidate region and select a candidate region with weak similarity as an object region.

• The KIPS Transactions:PartB
• /
• v.19B no.4
• /
• pp.243-248
• /
• 2012

In the information retrieval systems like vector model implementation and document clustering, document similarity calculation takes a great part on the overall performance of the system. In this paper, GPU parallelism has been explored to enhance the processing speed of document similarity calculation in a CUDA framework. The proposed method increased the similarity calculation speed almost 15 times better compared to the typical CPU-based framework. It is 5.2 and 3.4 times better than the methods by using CUBLAS and Thrust, respectively.