• The Mathematical Education
• /
• v.39 no.2
• /
• pp.167-177
• /
• 2000

In this thesis we concerned with the degree of similarity in pairs of scores having a common mean and variance could express similarity in terms of the absolute difference between the standardized scores. We particulary discussed the distribution of absolute differences between pairs of T scores among many standardized scores and demonstrated the procedures for calculating the lower limit of Med(│d│) values i. e. the maximum possible similarity with medians and correlation coefficients of the equivalent form tests by using the folded normal distribution, although other researchers expressed the degree of similarity using only the standard normal distribution. We also described many cases how to use those techniques and to apply effectively them in real evaluation fields.

• Proceedings of the Korean Society of Surveying, Geodesy, Photogrammetry, and Cartography Conference
• /
• 2006.04a
• /
• pp.243-248
• /
• 2006

In this paper, we propose the unsupervised change detection algorithm that apply the similarity measure techniques to the hyperspectral image. The general similarity measures including euclidean distance and spectral angle were compared. The spectral similarity scale algorithm for reducing the problems of those techniques was studied and tested with Hyperion data. The thresholds for detecting the change area were estimated through EM(Expectation-Maximization) algorithm. The experimental result shows that the similarity measure techniques and EM algorithm can be applied effectively for the unsupervised change detection of the hyperspectral data.

• Journal of the Korean Society for Aviation and Aeronautics
• /
• v.19 no.2
• /
• pp.16-22
• /
• 2011

It is recognized that the control surface fault is detected by monitoring the value of the coefficients due to the control surface deviation. It is found out the control surface stuck position by comparing the trim value with the reference value. To detect and isolate the fault, two mixed methods apply to the real-time parameter estimation and similarity measure. If the scatter of aerodynamic coefficients for the fault and normal are closing nearly, fault decision is difficult. Applying similarity measure to decide for fault or not, it makes a clear and easy distinction between fault and normal. Low power processor is applied to the real-time parameter estimator and computation of similarity measure.

• Phonetics and Speech Sciences
• /
• v.7 no.4
• /
• pp.49-55
• /
• 2015

It is argued in this paper that, in determining semantic similarity, Korean words should be recategorized with a focus on the semantic relation to ontology in light of cross-linguistic morphological variations. It is proposed, in particular, that Korean semantic similarity should be measured on three tracks, human judgements track, relatedness track, and cross-part-of-speech relations track. As demonstrated in Yang et al. (2015), GloVe, the unsupervised learning machine on semantic similarity, is applicable to Korean with its performance being compared with human judgement results. Based on this compatability, it was further thought that the model's performance might most likely vary with different kinds of specific relations in different languages. An attempt was made to analyze them in terms of two major Korean-specific categories involved in their lexical and cross-POS-relations. It is concluded that languages must be analyzed by varying methods so that semantic components across languages may allow varying semantic distance in the vector space models.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.11 no.7
• /
• pp.579-583
• /
• 2001

In the area of data clustering in high dimensional space, one of the difficulties is the time-consuming process for computing vector similarities. It becomes worse in the case of the agglomerative algorithm with the group-average link and mean centroid method, because the cluster similarity must be recomputed whenever the cluster center moves after the merging step. As a solution of this problem, we present an incremental method of similarity computation, which substitutes the scalar calculation for the time-consuming calculation of vector similarity with several measures such as the squared distance, inner product, cosine, and minimum variance. Experimental results show that it makes clustering speed significantly fast for very high dimensional data.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.7 no.2
• /
• pp.66-69
• /
• 1997

In this paper a fuzzy similarity measure is proposed. The proposed fuzzy similarity measure considers the relative distance between data and cluster centers in addition to the Euclidean distance to decide the degree of similarity. The boundary of a cluster center is constracted on the competitive region and expanded on the less competitive region. This result shows the possibility of using relative distance as a similarity measure.

• ETRI Journal
• /
• v.38 no.3
• /
• pp.540-550
• /
• 2016

Spectral clustering is a powerful tool for exploratory data analysis. Many existing spectral clustering algorithms typically measure the similarity by using a Gaussian kernel function or an undirected k-nearest neighbor (kNN) graph, which cannot reveal the real clusters when the data are not well separated. In this paper, to improve the spectral clustering, we consider a robust similarity measure based on the shared nearest neighbors in a directed kNN graph. We propose two novel algorithms for spectral clustering: one based on the number of shared nearest neighbors, and one based on their closeness. The proposed algorithms are able to explore the underlying similarity relationships between data points, and are robust to datasets that are not well separated. Moreover, the proposed algorithms have only one parameter, k. We evaluated the proposed algorithms using synthetic and real-world datasets. The experimental results demonstrate that the proposed algorithms not only achieve a good level of performance, they also outperform the traditional spectral clustering algorithms.

• Advanced Composite Materials
• /
• v.18 no.2
• /
• pp.135-152
• /
• 2009

Liquid molding processes, such as resin transfer molding, involve resin flow through a porous medium inside a mold cavity. Numerical analysis of resin flow and mold filling is a very useful means for optimization of the manufacturing process. However, the numerical analysis is quite time consuming and requires a great deal of effort, since a separate numerical calculation is needed for every set of material properties, part size and injection conditions. The efforts can be appreciably reduced if similarity solutions are used instead of repeated numerical calculations. In this study, the similarity relations for pressure, resin velocity and flow front propagation are proposed to correlate another desired case from the already obtained numerical result. In other words, the model gives a correlation of flow induced variables between two different cases. The model was verified by comparing results obtained by the similarity relation and by independent numerical simulation.

• Journal of the Korean Society of Visualization
• /
• v.6 no.2
• /
• pp.28-32
• /
• 2008

Contrary to the flow rate in fully filled pipe flows, the flow rate in partially filled pipe flows is significantly influenced by the variation of water level, channel slop, and so on. The major difference in these two flows results from the existence of a free surface. To make it clear, in the present study, a similarity of the velocity profile in a partially filled circular pipe has been investigated according to the water level. A particle image velocimetry (PIV) technique was applied to measure the three-dimensional velocity profiles. As a result, there is found a similarity of the velocity profile near the central region. However, near the side wall, the similarity is broken due to the interaction between the wall and the free surface.

• Proceedings of the KIEE Conference
• /
• 2006.04a
• /
• pp.159-161
• /
• 2006

The peer-to-peer (P2P) systems have Brown significantly over last few years due to their hish potential of sharing various resources. Super-peer based P2P systems have been found very effective by dividing the peers into two layers, SP (Super-Peer) and OP (Ordinary-Peer). In this paper, we present ISP2P (Intelligent Super-peer based P2P system), which allows us to choose the best SP. Through analyzing capacity and similarity between SP and OP, we can help OPs to select the most appropriate SP respectively. Proposed system can improve the performance of the average response time by superior SP, reduce the bandwidth cost by small path length due to content similarity and solve frequent SP replacement problem by considering similarity of user behavior.