• 한국통계학회:학술대회논문집
• /
• 한국통계학회 2005년도 추계 학술발표회 논문집
• /
• pp.181-186
• /
• 2005

Many data sets obtained from surveys or medical trials often include missing observations. When these data sets are analyzed, it is general to use only complete cases. However, it is possible to have big biases or involve inefficiency. In this paper, we consider a method for estimating parameters in logistic linear models involving non-ignorable missing data mechanism. A binomial response and normal exploratory model for the missing data are used. We fit the model using the EM algorithm. The E-step is derived by Metropolis-hastings algorithm to generate a sample for missing data and Monte-carlo technique, and the M-step is by Newton-Raphson to maximize likelihood function. Asymptotic variances of the MLE's are derived and the standard error and estimates of parameters are compared.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2004년도 ICEIC The International Conference on Electronics Informations and Communications
• /
• pp.438-441
• /
• 2004

Image binarization using a global threshold value [3] performs at high speed, but usually results in undesired binary images when the source images are of poor quality. In such cases, adaptive local thresholding algorithms [1][2][3] are used to obtain better results, and the algorithm proposed by A.E.Savekis which chooses local threshold using fore­ground and background clustering [1] is one of the best thresholding algorithms. However, this algorithm runs slowly due to its re-computing threshold value of each central pixel in a local window MxM. In this paper, we present a dynamic programming approach for the step of calculating local threshold value that reduces many redundant computations and improves the execution speed significantly. Experiments show that our proposal improvement runs more ten times faster than the original algorithm.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2014년도 춘계학술발표대회
• /
• pp.830-833
• /
• 2014

This paper proposes dynamic constraint parameter to filter out degenerate configurations (i.e. set of collinear or adjacent features) in RANSAC algorithm. We define five different groups of image based on the feature distribution pattern. We apply the same linear and distance constraints for every image, but we use different constraint parameter for every group, which will affect the filtering result. An evaluation is done by comparing the proposed dynamic CS-RANSAC algorithm with the classic RANSAC and regular CS-RANSAC algorithms in the calculation of a homography matrix. The experimental results show that dynamic CS-RANSAC algorithm provides the lowest error rate compared to the other two algorithms.

• Communications for Statistical Applications and Methods
• /
• 제29권3호
• /
• pp.333-351
• /
• 2022

The skew-t distribution is an attractive family of asymmetrical heavy-tailed densities that includes the normal, skew-normal and Student's-t distributions as special cases. In this work, we propose an EM-type algorithm for computing the maximum likelihood estimates for skew-t linear regression models with censored response. In contrast with previous proposals, this algorithm uses analytical expressions at the E-step, as opposed to Monte Carlo simulations. These expressions rely on formulas for the mean and variance of a truncated skew-t distribution, and can be computed using the R library MomTrunc. The standard errors, the prediction of unobserved values of the response and the log-likelihood function are obtained as a by-product. The proposed methodology is illustrated through the analyses of simulated and a real data application on Letter-Name Fluency test in Peruvian students.

• 한국정보처리학회논문지
• /
• 제3권1호
• /
• pp.43-54
• /
• 1996

이 논문은 새로운 휴리스틱 탐색(heuristic search)방법을 이용하여, 수평 및 수 직선으로 이루어진 방해 물들이 놓인 가운데 수평 및 수직선으로 구성된 최단 거리 (rectilinear shortestpath)와 꺾이는회수가 가장 적은최소 꺾임경로(link metric shortest path) 및 이 둘을 혼합시킨 혼합형 최단 경로를 구하는 알고리즘을 서술 하고 있다. 최단 경로를 구하는 방법으로 미로 찾기형 알고리즘(maze-running algorithms)과 선형 탐색 알고리즘(line-search algorithms)의 장점만을 이용한 GMD 알고리즘(Guided Minimum Detour algorithm)을 제안하고 있으며 이를 더욱 효율 적으 로 개선한 LGMD 알고리즘 (Line-by-Line Guided Minimum Detour algorithmm)을 개발 하였다. 이들 GMD와 LGMD 알고리즘은 기존의 최단 경로를 내포하고 있는 conection group를 이용하지 않고서도 휴리스틱을 사용한 guided A 탐색(guided A* search)을 이용하여 최적의 최단 경로를 구할 수 있는 장점이 있으며 시간과 메모리 면에서 효 율을 극대화하였다. 이들 GMD와 LGMD 알고리즘은 각각 O(m＋eloge＋NlogN)와 O(eloge＋ NlogN)의 시간과 O(e＋N)의 메모리를 사용한다. 여기서 m은 탐색에 사용된 지선 (line segment)들의 수이다. 또한 LGMD는 최소 꺾임 경로(link metric shortest path)와 최단 경로와 최소의 꺾임을 조합한 혼합형 최단 경로를 구하는 데에도 적용될 수 있는 확장성을 가지고 있다.

• Journal of the Korean Society for Industrial and Applied Mathematics
• /
• 제20권2호
• /
• pp.107-121
• /
• 2016

We analyze the complexity of the LLL algorithm, invented by Lenstra, Lenstra, and $Lov{\acute{a}}sz$ as a a well-known lattice reduction (LR) algorithm which is previously known as having the complexity of $O(N^4{\log}B)$ multiplications (or, $O(N^5({\log}B)^2)$ bit operations) for a lattice basis matrix $H({\in}{\mathbb{R}}^{M{\times}N})$ where B is the maximum value among the squared norm of columns of H. This implies that the complexity of the lattice reduction algorithm depends only on the matrix size and the lattice basis norm. However, the matrix structures (i.e., the correlation among the columns) of a given lattice matrix, which is usually measured by its condition number or determinant, can affect the computational complexity of the LR algorithm. In this paper, to see how the matrix structures can affect the LLL algorithm's complexity, we derive a more tight upper bound on the complexity of LLL algorithm in terms of the condition number and determinant of a given lattice matrix. We also analyze the complexities of the LLL updating/downdating schemes using the proposed upper bound.

• Journal of Electrical Engineering and Technology
• /
• 제6권2호
• /
• pp.182-191
• /
• 2011

Genetic algorithms (GA) and particle swarm optimization (PSO) are the most famous optimization techniques among various modern heuristic optimization techniques. These two approaches identify the solution to a given objective function, but they employ different strategies and computational effort; therefore, a comparison of their performance is needed. This paper presents the application and performance comparison of the PSO and GA optimization techniques for a static synchronous series compensator-based controller design. The design objective is to enhance power system stability. The design problem of the FACTS-based controller is formulated as an optimization problem, and both PSO and GA optimization techniques are employed to search for the optimal controller parameters.

• 대한산업공학회지
• /
• 제23권3호
• /
• pp.487-500
• /
• 1997

In this paper, we introduce the shortest flow-generating path problem in the generalized network. As the simplest generalized network model, this problem captures many of the most salient core ingredients of the generalized network flows and so it provides both a benchmark and a point of departure for studying more complex generalized network models. We show that the generalized label-correcting algorithm for the shortest flow-generating path problem has O(mn) time complexity if it starts with a good point and also propose an O($n^3m^2$) algorithm for finding a good starting point. Hence, the shortest flow-generating path problem is solved in O($n^3m^2$) time.

• Journal of Electrical Engineering and Technology
• /
• 제12권5호
• /
• pp.1980-1990
• /
• 2017

A study on the multi-objective optimization of Interior Permanent-Magnet Synchronous Motors (IPMSMs) with 2, 3, 4 and 5 flux barriers per magnetic pole, based on Genetic Algorithm (GA) is presented by considering the aspect of irreversible demagnetization. Applying the 2004 Toyota Prius single-layer IPMSM as the reference machine, the asymmetrical two-, three-, four- and five-layer rotor models with the same amount of Permanent-Magnets (PMs) is presented to improve the torque characteristics, i.e., reducing the torque pulsation and increasing the average torque. A reduction of the torque pulsations is achieved by adopting different and asymmetrical flux barrier geometries in each magnetic pole of the rotor topology. The demagnetization performance in the PMs is considered as well as the motor performance; and analyzed by using finite element method (FEM) for verification of the optimal solutions.

• Communications for Statistical Applications and Methods
• /
• 제9권1호
• /
• pp.155-166
• /
• 2002

Neural networks have been studied as a popular tool for classification and they are very flexible. Also, they are used for many applications of pattern classification and pattern recognition. This paper focuses on Bayesian approach to feed-forward neural networks with single hidden layer of units with logistic activation. In this model, we are interested in deciding the number of nodes of neural network model with p input units, one hidden layer with m hidden nodes and one output unit in Bayesian setup for fixed m. Here, we use the latent variable into the prior of the coefficient regression, and we introduce the 'sequential step' which is based on the idea of the data augmentation by Tanner and Wong(1787). The MCMC method(Gibbs sampler and Metropolish algorithm) can be used to overcome the complicated Bayesian computation. Finally, a proposed method is applied to a simulated data.