• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.9
• /
• pp.2991-3007
• /
• 2022

Two dimensional locality preserving projections (2D-LPP) is an improved algorithm of 2D image to solve the small sample size (SSS) problems which locality preserving projections (LPP) meets. It's able to find the low dimension manifold mapping that not only preserves local information but also detects manifold embedded in original data spaces. However, 2D-LPP is simple and elegant. So, inspired by the comparison experiments between two dimensional linear discriminant analysis (2D-LDA) and linear discriminant analysis (LDA) which indicated that matrix based methods don't always perform better even when training samples are limited, we surmise 2D-LPP may meet the same limitation as 2D-LDA and propose a novel matrix exponential method to enhance the performance of 2D-LPP. 2D-MELPP is equivalent to employing distance diffusion mapping to transform original images into a new space, and margins between labels are broadened, which is beneficial for solving classification problems. Nonetheless, the computational time complexity of 2D-MELPP is extremely high. In this paper, we replace some of matrix multiplications with multiple multiplications to save the memory cost and provide an efficient way for solving 2D-MELPP. We test it on public databases: random 3D data set, ORL, AR face database and Polyu Palmprint database and compare it with other 2D methods like 2D-LDA, 2D-LPP and 1D methods like LPP and exponential locality preserving projections (ELPP), finding it outperforms than others in recognition accuracy. We also compare different dimensions of projection vector and record the cost time on the ORL, AR face database and Polyu Palmprint database. The experiment results above proves that our advanced algorithm has a better performance on 3 independent public databases.

• Journal of Positioning, Navigation, and Timing
• /
• v.4 no.2
• /
• pp.79-85
• /
• 2015

Existing Differential GPS (DGPS) pseudorange correction (PRC) generation techniques based on a virtual reference station cannot effectively assign a weighting factor if the baseline distance between a user and a reference station is not long enough. In this study, a virtual reference station DGPS PRC generation technique was developed based on an enhanced inverse distance weighting method using an exponential function that can maximize a small baseline distance difference due to the dense arrangement of DGPS reference stations in South Korea, and its positioning performance was validated. For the performance verification, the performance of the model developed in this study (EIDW) was compared with those of typical inverse distance weighting (IDW), first- and second-order multiple linear regression analyses (Planar 1 and 2), the model of Abousalem (1996) (Ab_EXP), and the model of Kim (2013) (Kim_EXP). The model developed in the present study had a horizontal accuracy of 53 cm, and the positioning based on the second-order multiple linear regression analysis that showed the highest positioning accuracy among the existing models had a horizontal accuracy of 51 cm, indicating that they have similar levels of performance. Also, when positioning was performed using five reference stations, the horizontal accuracy of the developed model improved by 8 ~ 42% compared to those of the existing models. In particular, the bias was improved by up to 27 cm.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.29 no.1C
• /
• pp.129-139
• /
• 2004

This paper studies the support region of an optimum (minimum mean-squre error) fixed-rate scalar quantizer for a Weibull source. The support region is defined to be the interval determined by the outermost thresholds of a quantizer and plays an important role in its performance, and hence it motivates this study. The paper reports the following specific results. First, approximation formulas are derived for the outermost thresholds of optimum scalar quantizers for a Weibull distributions. Second, in the case of Rayleigh and exponential distributions the derived approximation formulas are compared for the evaluation of their accuracy with the true values of optimum quantizers. Numerical results show that the formula for the leftmost threshold stays within 1% of the true value for 128 and 256 quantization points or more, for Rayleigh and exponential distribution, respectively, while that for the rightmost threshold does so for 512 and 32 quantization points or more. These formulas exhibit increased accuracy with the number of quantization points. In conclusion, the formulas have high accuracy. The contribution of the paper consists in the derivation of closed accurate formulas for the support of optimum.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.3 no.4
• /
• pp.251-259
• /
• 2008

This paper presents the hardware design of a 32bit floating point based processor. The processor can perform nonlinear functions such as sinusoidal functions, exponential functions, and other mathematical functions. Using the Taylor series and Newton - Raphson method, nonlinear functions are approximated. The processor is actually embedded on an FPGA chip and tested. The numerical accuracy of the functions is compared with those computed by the MATLAB and confirmed the performance of the processor.

• Journal of applied mathematics & informatics
• /
• v.32 no.5_6
• /
• pp.761-781
• /
• 2014

An exponentially accurate nonconforming spectral element method for elasticity systems with discontinuities in the coefficients and the flux across the interface is proposed in this paper. The method is least-squares spectral element method. The jump in the flux across the interface is incorporated (in appropriate Sobolev norm) in the functional to be minimized. The interface is resolved exactly using blending elements. The solution is obtained by the preconditioned conjugate gradient method. The numerical solution for different examples with discontinuous coefficients and non-homogeneous jump in the flux across the interface are presented to show the efficiency of the proposed method.

• Journal of the Korean Data and Information Science Society
• /
• v.23 no.6
• /
• pp.1279-1287
• /
• 2012

We particularly emphasized how to determine the number of replications with sequential and multistage procedures. So, the t-test is used to achieve some predetermined level of accuracy efficiently with loss function in the case of normal, chi-squared, an exponential distributions. We provided that the relevance of procedures are sequential procedure, two-stage procedure, modified two-stage procedure, three-stage procedure and accelerated sequential procedure. Monte Carlo simulation is carried out to obtain the stopping sample size that minimizes the risk.

• Proceedings of the KIEE Conference
• /
• 2007.10a
• /
• pp.74-76
• /
• 2007

This paper presents the hardware design of a 32bit floating point based processor. The processor can perform nonlinear functions such as sinusoidal functions, exponential functions, and other nonlinear functions. Using the Taylor series and the Newton - Raphson method, nonlinear functions are approximated. The processor is actually embedded on an FPGA chip and tested. The numerical accuracy of the functions is compared with those computed by the MATLAB.

• Journal of the Korean Data and Information Science Society
• /
• v.17 no.3
• /
• pp.687-705
• /
• 2006

In this paper we consider GPH distribution that is defined as a distribution for sum of random number of random variables following exponential distribution. We establish approximation process of general distributions to GPH distributions and offer numerical results for various cases to show the accuracy of the approximation. We also propose analysis method of delay distribution of queueing systems using approximation to GPH distributions and offer numerical results for various queueing systems to show applicability of GPH approximation.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.31 no.4
• /
• pp.490-495
• /
• 2007

This research aims to examine accuracy and efficiency of the first four moments corresponding to mean, standard deviation, skewness, and kurtosis, which are estimated by a function approximation moment method (FAMM). In FAMM, the moments are estimated from an approximating quadratic function of a system response function. The function approximation is performed on a specially selected experimental region for accuracy, and the number of function evaluations is taken equal to that of the unknown coefficients for efficiency. For this purpose, three error-minimizing conditions are utilized and corresponding canonical experimental regions constructed accordingly. An interpolation function is then obtained using a D-optimal design and then the first four moments of it are obtained as the estimates for the system response function. In order to verify accuracy and efficiency of FAMM, several non-linear examples are considered including a polynomial of order 4, an exponential function, and a rational function. The moments calculated from various coefficients of variation show very good accuracy and efficiency in comparison with those from analytic integration or the Monte Carlo simulation and the experimental design technique proposed by Taguchi and updated by D'Errico and Zaino.

• IE interfaces
• /
• v.11 no.3
• /
• pp.55-63
• /
• 1998

In this study we consider a CONWIP system in which the processing times at each station follow an exponential distribution and the demands for the finished products arrive according to a compound Poisson process. The demands that are not satisfied instantaneously are assumed to be lost. We assume that the lot size at each station is greater than one. For this system we develop an approximation method to obtain the performance measures such as steady state probabilities of the number of parts at each station, average number of parts at each station and the proportion of lost demands. For the analysis of the proposed CONWIP system, we model the CONWIP system as a closed queueing network with a synchronization station and analyze the closed queueing network using a product form approximation method. A recursive technique is used to solve the subnetwork in the application of the product-form approximation method. To test the accuracy of the approximation method, the results obtained from the approximation method were compared with those obtained by simulation. Comparisons with simulation have shown that the accuracy of the approximate method is acceptable.