• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.9 no.2
• /
• pp.742-762
• /
• 2015

In this study, we propose a distance metric learning approach called discriminant metric learning (DML) for face verification, which addresses a binary-class problem for classifying whether or not two input images are of the same subject. The critical issue for solving this problem is determining the method to be used for measuring the distance between two images. Among various methods, the large margin nearest neighbor (LMNN) method is a state-of-the-art algorithm. However, to compensate the LMNN's entangled data distribution due to high levels of appearance variations in unconstrained environments, DML's goal is to penalize violations of the negative pair distance relationship, i.e., the images with different labels, while being integrated with LMNN to model the distance relation between positive pairs, i.e., the images with the same label. The likelihoods of the input images, estimated using DML and LMNN metrics, are then weighted and combined for further analysis. Additionally, rather than using the k-nearest neighbor (k-NN) classification mechanism, we propose a verification mechanism that measures the correlation of the class label distribution of neighbors to reduce the false negative rate of positive pairs. From the experimental results, we see that DML can modify the relation of negative pairs in the original LMNN space and compensate for LMNN's performance on faces with large variances, such as pose and expression.

• Journal of the Korean Data and Information Science Society
• /
• v.20 no.1
• /
• pp.109-115
• /
• 2009

Many internet companies are holding marketing activity through customer relationship management to satisfy complicated and diversified consumer demands. Use of customer lifetime value as a marketing metric tends to place greater emphasis on customer service and long-term customer satisfaction, rather than on maximizing short term sales. And so many internet companies have been interested in customer lifetime value, which is a primary key for discovery customer values to promote the competitive power in their business fields. In this paper, we apply a life table technique to lifetime analysis of internet site customers by membership pattern and provide the opportunity using revised life tables in several kinds of internet companies.

• The Transactions of the Korean Institute of Electrical Engineers D
• /
• v.49 no.8
• /
• pp.491-494
• /
• 2000

A scheme is proposed for measuring similarities between the binary pulse signals in the pulse-code modulation using the Integra-Normalizer. The Integra-Normalizer provides a better interpretation of the relationship between the pulse signals by removing redundant codes, which maps all possible observed signals to one of the codes to be received with relative similarities between each pair of compared signals. The proposed method provides better error tolerance than L2 metric, such as Hamming distance, since the distances between pulse signals are measured not useful for the time-delay detection in the pulse-code modulation.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.6 no.6
• /
• pp.1649-1667
• /
• 2012

In order to improve the accuracy and universality of the flow metric correlation analysis, this paper firstly analyzes the characteristics of Internet flow metrics as random variables, points out the disadvantages of Pearson Correlation Coefficient which is used to measure the correlation between two flow metrics by current researches. Then a method based on Symmetrical Uncertainty is proposed to measure the correlation between two flow metrics, and is extended to measure the correlation among multi-variables. Meanwhile, the simulation and polynomial fitting method are used to reveal the threshold value between different correlation degrees for SU method. The statistical analysis results on the common flow metrics using several traces show that Symmetrical Uncertainty can not only represent the correct aspects of Pearson Correlation Coefficient, but also make up for its shortcomings, thus achieve the purpose of measuring flow metric correlation quantitatively and accurately. On the other hand, reveal the actual relationship among fourteen common flow metrics.

• The Korean Journal of Applied Statistics
• /
• v.28 no.6
• /
• pp.1171-1180
• /
• 2015

Multidimensional scaling (MDS) is a graphical technique of multivariate analysis to display dissimilarities among individuals into low-dimensional space. We often have two kinds of MDS which are metric MDS and non-metric MDS. Metric MDS can be applied to quantitative data; however, we need additional information about variables because it only shows relationships among individuals. Gower (1992) proposed a method that can represent variable information using trajectories of the pseudo-points for quantitative variables on the metric MDS space. We will call his method a 'replacement method'. However, the trajectory can not be represented even though metric MDS can be applied to binary data when we apply his method to binary data. Therefore, we propose a method to represent information of binary variables using pseudo-points called a 'partition method'. The proposed method partitions pseudo-points, accounting both the rate of zeroes and ones. Our metric MDS using the proposed partition method can show the relationship between individuals and variables for binary data.

• Journal of Broadcast Engineering
• /
• v.19 no.1
• /
• pp.31-43
• /
• 2014

A depth map is an important component for stereoscopic image generation. Since the depth map acquired from a depth camera has a low resolution, upsamling a low-resolution depth map to a high-resolution one has been studied past decades. Upsampling methods are evaluated by objective evaluation tools such as PSNR, Sharpness Degree, Blur Metric. As well, the subjective quality is compared using virtual views generated by DIBR (depth image based rendering). However, works on the analysis of the relation between depth map upsampling and stereoscopic images are relatively few. In this paper, we investigate the relationship between subjective evaluation of stereoscopic images and objective performance of upsampling methods using cross correlation and linear regression. Experimental results demonstrate that the correlation of edge PSNR and visual fatigue is the highest and the blur metric has lowest correlation. Further, from the linear regression, we found relative weights of objective measurements. Further we introduce a formulae that can estimate 3D performance of conventional or new upsampling methods.

• Journal of Korean Society of Forest Science
• /
• v.21 no.1
• /
• pp.47-51
• /
• 1974

In order to prove the similarity relationships between the basic species of oak through Electronic Data Processing System(EDPS) and numerical analysis, The analized species and datas were selected from the list of morphological observation in the thesis of T.B. Lee, 1961, "Phytogenetic study of the subgenus Lepidobalanus of the genus Quercus in Korea", and were coded by categories shown in Table 1. The value in the list were transformed into hundred percentage to standardize the observational value by each code into dimensionless. The similarity index between species were computed through formula of non-metric coefficient, $N_{jk}=\sum\limits_{i=1}^{n}\(\frac{{\mid}x_{ij}-x_{ik}{\mid}}{x_{ij}+x_{ik}}\)$, using the UNiVAC-1106, at National Computer Center. Quercus aliena, by analysis result, is most similar to Q. mongolica with the similarity index, 71.6 and Q. dentata is most far apart from Q. serrata in the relationship with index, 121.4. The above thesis of Professor, T. Lee, are closely similar with the result of this research study. But, their similar relationship are proved in quantity through numerical method in our research study. In addition, The relationships among Q. mongolica, Q. aliena and Q. serrata are found to be very similar, but Q. dentata to be enough far in similarity to other species by dendrogram shown at Fig. 1. The numerical classification through EDPS is found to be suitable method also applicable to the plant taxonomy.

• Journal of Digital Contents Society
• /
• v.12 no.3
• /
• pp.339-345
• /
• 2011

The k-means has been one of the popular, simple and faster clustering algorithms, but the right value of k is unknown. The value of k (the number of clusters) is a very important element because the result of clustering is different depending on it. In this paper, we present a novel algorithm based on an external cluster relationship criterion which is an evaluation metric of clustering result to determine the number of clusters dynamically. Experimental results show that our algorithm is superior to other methods in terms of the accuracy of the number of clusters.

• International Journal of Reliability and Applications
• /
• v.4 no.3
• /
• pp.131-139
• /
• 2003

There is a new trend of incorporating software coverage metrics into software modelling. This paper proposes and empirically evaluates a software reliability growth model, which relates reliability to coverage. The proposed model is derived by modifying the assumptions on which Veevers and Marshall model is based.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.9 no.5
• /
• pp.1234-1241
• /
• 2008

Cost effectiveness is greatly related with the degree of reducing the testing cost by the technical reviews. In this paper, we present a new metric My for evaluating the cost effectiveness of technical reviews during software development. First, we estimate and compare My with conventional measure using data collected during practical software development procedure, then we show the validity and usefulness of the proposed measure My. Also by formulating the relationship between the data collected during the reviews and the test, we present a method to estimate the value of the metric My using only the data collected during review phase.