• The Journal of the Korea Contents Association
• /
• v.20 no.6
• /
• pp.556-578
• /
• 2020

There are many different researches about how to analyze issues based on real-time news streams. But, there are few researches which analyze issues hierarchically from news articles and even a previous research of hierarchical issue analysis make clustering speed slower as the increment of news articles. In this paper, we propose a hierarchical and incremental clustering for semi real-time issue analysis on news articles. We trained siamese neural network based weighted cosine similarity model, applied this model to k-means algorithm which is used to make word clusters and converted news articles to document vectors by using these word clusters. Finally, we initialized an issue cluster tree from document vectors, updated this tree whenever news articles happen, and analyzed issues in semi real-time. Through the experiment and evaluation, we showed that up to about 0.26 performance has been improved in terms of NMI. Also, in terms of speed of incremental clustering, we also showed about 10 times faster than before.

• Journal of Intelligence and Information Systems
• /
• v.10 no.2
• /
• pp.15-37
• /
• 2004

With the growth in the size of datasets, data mining has recently become an important research topic. Especially, interests about spatio-temporal data mining has been increased which is a method for analyzing massive spatio-temporal data collected from a wide variety of applications like GPS data, trajectory data of surveillance system and earth geographic data. In the former approaches, conventional clustering algorithms are applied as spatio-temporal data mining techniques without any modification. In this paper, we focused to SOM that is the most common clustering algorithm applied to clustering analysis in data mining wet and develop the spatio-temporal data mining module based on it. In addition, we analyzed the clustering results of developed SOM module and compare them with those of K-means and Agglomerative Hierarchical algorithm in the aspects of homogeneity, separation, separation, silhouette width and accuracy. We also developed specialized visualization module fur more accurate interpretation of mining result.

• Journal of Korea Water Resources Association
• /
• v.51 no.11
• /
• pp.959-969
• /
• 2018

The aging water pipes induce various problems, such as water supply suspension due to breakage, insufficient water pressure, deterioration of water quality, damage by sink holes, and economic losses due to water leaks. However, it is impractical and almost impossible to repair and/or replace all deteriorated water pipes simultaneously. Hence, it is required to quantitatively evaluate the deterioration rate of individual pipes indirect way to determine the rehabilitation order of priority. In this study, ANN(Artificial Neural Network)-Clustering method is suggested as a new approach to assess and assort the water pipes. The proposed method has been applied to a water supply network of YG-county in Jeollanam-do. To assess the applicability of the model, the evaluation results were compared with the results of the Numerical Weighting Method (NWM), which is being currently utilized in practice. The assessment results are depicted in a water pipe map to intuitively grasp the degree of deterioration of the entire pipelines. The application results revealed that the proposed ANN-Clustering models can successfully assess the water pipe deterioration along with the conventional approach of NWM.

• Journal of Digital Convergence
• /
• v.18 no.3
• /
• pp.189-194
• /
• 2020

Recently, various clustering techniques have been studied to efficiently handle data generated by heterogeneous IoT devices. However, existing clustering techniques are not suitable for mobile IoT devices because they focus on statically dividing networks. This paper proposes a probabilistic deep learning-based dynamic clustering model for collecting and analyzing information on IoT devices using edge networks. The proposed model establishes a subnet by applying the frequency of the attribute values collected probabilistically to deep learning. The established subnets are used to group information extracted from seeds into hierarchical structures and improve the speed and accuracy of dynamic clustering for IoT devices. The performance evaluation results showed that the proposed model had an average 13.8 percent improvement in data processing time compared to the existing model, and the server's overhead was 10.5 percent lower on average than the existing model. The accuracy of extracting IoT information from servers has improved by 8.7% on average from previous models.

• The KIPS Transactions:PartD
• /
• v.11D no.3
• /
• pp.717-730
• /
• 2004

Recently. many researches on the personalization of a web-site have been actively made. The web personalization predicts the sets of the most interesting URLs for each user through data mining approaches such as clustering techniques. Most existing methods using clustering techniques represented the web transactions as bit vectors that represent whether users visit a certain WRL or not to cluster web transactions. The similarity of the web transactions was decided according to the match degree of bit vectors. However, since the existing methods consider only whether users visit a certain URL or not, users' interestingness on the URL is excluded from clustering web transactions. That is, it is possible that the web transactions with different visit proposes or inclinations are classified into the same group. In this paper. we propose an enhanced transaction modeling with interestingness weight to solve such problems and a new similarity measuring method that exploits the proposed transaction modeling. It is shown through performance evaluation that our similarity measuring method improves the accuracy of the web transaction clustering over the existing method.

• Journal of Digital Contents Society
• /
• v.14 no.1
• /
• pp.51-57
• /
• 2013

Storage device based on Flash Memory have many attractive features such as high performance, low power consumption, shock resistance, and low weight, so they replace HDDs to a certain extent. An Storage device based on Flash Memory has FTL(Flash Translation Layer) which emulate block storage devices like HDDs. A garbage collection, one of major functions of FTL, effects highly on the performance and the lifetime of devices. However, there is no de facto standard for new garbage collection algorithms. To solve this problem, we propose File Clustering Algorithm. File Clustering Algorithm respect to update page from same file at the same time. So, these are clustered to same block. For this mechanism, We propose Page Allocation Policy in FTL and use MIN-MAX GAP to guarantee wear leveling. To verify the algorithm in this paper, we use TPC Benchmark. So, The performance evaluation reveals that the proposed algorithm has comparable result with the existing algorithms(No wear leveling, Hot/Cold) and shows approximately 690% improvement in terms of the wear leveling.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.12 no.2
• /
• pp.151-156
• /
• 2002

There are two methods in evaluation student s achievement. The two evaluation methods are absolute evaluation and relative evaluation. They have much advantages respectively, but also have some limitations such as being too stereotyped or causing overcompetition among learners. This paper suggests a new evaluation method which evaluates student s achievements by considering the score distribution and the frequency The proposed method classifies the scores into several clusters considering the goodness. This approach calculates the goodness by applying the RE(relaxation error), and grades the achievement scores based on the goodness. The suggested method can avoid the problem of grading caused by the narrow gap of scores because it sets a standard for grading by the calculated goodness considering the score distribution and frequency of occurrence. The method can differentiate achievements of a school from those of others, and that it is useful for selecting advanced students and dull ones, and for evaluation of classes based on student s achievement.

• Asian Pacific Journal of Cancer Prevention
• /
• v.15 no.14
• /
• pp.5721-5727
• /
• 2014

Objectives: To evaluate the performance of clustering methods used in the prognostic assessment of categorical clinical data for hepatocellular carcinoma (HCC) patients in China, and establish a predictable prognostic nomogram for clinical decisions. Materials and Methods: A total of 332 newly diagnosed HCC patients treated with hepatic resection during 2006-2009 were enrolled. Patients were regularly followed up at outpatient clinics. Clustering methods including the Average linkage, k-modes, fuzzy k-modes, PAM, CLARA, protocluster, and ROCK were compared by Monte Carlo simulation, and the optimal method was applied to investigate the clustering pattern of the indices including platelet count, platelet/lymphocyte ratio (PLR) and serum aspartate aminotransferase activity/platelet count ratio index (APRI). Then the clustering variable, age group, tumor size, number of tumor and vascular invasion were studied in a multivariable Cox regression model. A prognostic nomogram was constructed for clinical decisions. Results: The ROCK was best in both the overlapping and non-overlapping cases performed to assess the prognostic value of platelet-based indices. Patients with categorical platelet-based indices significantly split across two clusters, and those with high values, had a high risk of HCC recurrence (hazard ratio [HR] 1.42, 95% CI 1.09-1.86; p<0.01). Tumor size, number of tumor and blood vessel invasion were also associated with high risk of HCC recurrence (all p< 0.01). The nomogram well predicted HCC patient survival at 3 and 5 years. Conclusions: A cluster of platelet-based indices combined with other clinical covariates could be used for prognosis evaluation in HCC.

• The KIPS Transactions:PartD
• /
• v.13D no.5 s.108
• /
• pp.633-640
• /
• 2006

Spatial clustering, which groups similar objects based on their distance, connectivity, or their relative density in space, is an important component of spatial data mining. Density-based and grid-based clustering are two main clustering approaches. The former is famous for its capability of discovering clusters of various shapes and eliminating noises, while the latter is well known for its high speed. Clustering large data sets has always been a serious challenge for clustering algorithms, because huge data set would make the clustering process extremely costly. In this paper, we propose an enhanced Density-Grid based Clustering algorithm for Large spatial database by setting a default number of intervals and removing the outliers effectively with the help of a proper measurement to identify areas of high density in the input data space. We use a density threshold DT to recognize dense cells before neighbor dense cells are combined to form clusters. When proposed algorithm is performed on large dataset, a proper granularity of each dimension in data space and a density threshold for recognizing dense areas can improve the performance of this algorithm. We combine grid-based and density-based methods together to not only increase the efficiency but also find clusters with arbitrary shape. Synthetic datasets are used for experimental evaluation which shows that proposed method has high performance and accuracy in the experiments.

• Journal of KIISE:Software and Applications
• /
• v.31 no.9
• /
• pp.1246-1257
• /
• 2004

To find an optimal solution with genetic algorithm, it is desirable to maintain the population sire as large as possible. In some cases, however, the cost to evaluate each individual is relatively high and it is difficult to maintain large population. To solve this problem we propose a novel genetic algorithm based on fuzzy clustering, which considerably reduces evaluation number without any significant loss of its performance by evaluating only one representative for each cluster. The fitness values of other individuals are estimated from the representative fitness values indirectly. We have used fuzzy c-means algorithm and distributed the fitness using membership matrix, since it is hard to distribute precise fitness values by hard clustering method to individuals which belong to multiple groups. Nine benchmark functions have been investigated and the results are compared to six hard clustering algorithms with Euclidean distance and Pearson correlation coefficients as fitness distribution method.