• 응용통계연구
• /
• 제28권2호
• /
• pp.353-360
• /
• 2015

공간적으로 관측되는 연속형 자료를 분석하는 모형으로 공간적 상관관계를 고려한 다양한 정규모형이 지난 수십 년간 제안되었다. 그 중에서 공간효과를 랜덤효과로 모형화하는 공간선형모형(Spatial Linear Mixed Model; SLMM)이 가장 널리 활용되는 모형 중 하나일 것이다. 연결함수(link function)을 사용하면 SLMM을 비정규 데이터도 적용할 수 있는 일반화된 공간선형모형(Spatial Generalized Linear Mixed Model; SGLMM)으로 자연스럽게 확장할 수 있다. 이 논문에서는 가장 널리 활용되는 SGLMM을 알아보고 실제 데이터 적용사례를 R 패키지를 활용하여 제시하고자 한다.

• Wind and Structures
• /
• 제2권2호
• /
• pp.85-94
• /
• 1999

The EARPG(1)/UPS was first developed by Seong (1993) and has been tested for wind pressure time series simulations (Seong and Peterka 1993, 1997, 1998) to prove its excellent performance for generating non-Gaussian time series, in particular, with large amplitude sharp peaks. This paper presents a parametric study focused on simulation of extreme value statistics based on the synthetic realizations of the EARPG(1)/UPS. The method is shown to have a great capability to simulate a wide range of non-Gaussian statistic values and extreme value statistics with exact target sample power spectrum. The variation of skewed long tail in PDF and extreme value distribution are illustrated as function of relevant parameters.

• 한국전산구조공학회논문집
• /
• 제13권4호
• /
• pp.427-436
• /
• 2000

이 논문은 불규칙 가진력을 받는 회전-고정보의 비선형 응답특성을 나타낸다. 불규칙 가진력은 두 번째 고유모드의 절점과 최대변위점에 가했다. 비선형 편미분 방정식과 경계조건으로 표현되는 이 문제를 Galerkin의 방법을 이용하여 연립 비선형 상미분방정식으로 변환하였다. 이 상미분방정식으로부터 Fokker-Planck방정식과 모멘트 방정식을 얻은 후 Gaussian closure 방법 및 non-Gaussian closure 방법을 이용하여 3 모드 근사시 각각 27개 및 209개의 자율 상미분방정식을 구하였다. Gaussian closure 방법과 non-Gaussian closure 방법으로 2 모드 및 3 모드 근사해석을 수행하였고 해석적 결과들은 Monte Carlo 시뮬레이션 결과와 비교되었다. 해석결과 2 모드 근사해와 3 모드 근사해가 거의 일치하였고 2 모드 내부공진만 고려하여도 해석결과에 별 영향을 주지 않는다는 것을 알 수 있다.

• 응용통계연구
• /
• 제27권1호
• /
• pp.31-41
• /
• 2014

행렬도(biplot)는 이원표 자료행렬(two-way data matrix)의 행과 열을 한 그림에 동시에 나타내는 탐색적 방법으로, 복잡한 다변량 분석 결과를 보다 쉽게 파악할 수 있는 장점이 있다. 특히 주성분인자 행렬도(principal component factor biplot; PCFB)는 인자분석을 통해서 변수들 간의 상호의존 구조를 탐색하기 위한 시각적 도구이다. 자료에 따라 잠재된 변수들이 독립(independent)이고 비가우시안(non-Gaussian) 분포를 가진다는 사전 정보가 있을 때, Jutten과 Herault (1991)가 제안한 독립성분분석(independent component analysis)을 이용한다. 이 경우 주성분법을 이용한 인자분석을 적용하면 원래 변수들의 상호 관계를 잘못 해석할 수도 있다. 따라서 본 논문에서는 자료에 따라 잠재된 변수들이 독립이고 비가우시안 분포를 가진다는 사전 정보가 있을 때, 독립성분분석을 응용하여 원래 변수들 간의 상호 관계를 기하학적으로 살펴볼 수 있는 시각적 도구인 독립성분 행렬도(independent component biplot; ICB)를 제안하려 한다.

• Wind and Structures
• /
• 제6권5호
• /
• pp.357-386
• /
• 2003

A number of methods based on various ideas have been proposed for simulating the non-Gaussian stationary process. However, these methods have some limitations. This paper reviewed several simulation methods based on the translation method using logarithmic and polynomial functions, which have emerged in the history of statistics and in the field of civil engineering. The applicability of each method is discussed from the viewpoint of the reproducibility of higher order statistics of the object function in the simulated sample functions, and examined using pressure signals measured from wind tunnel experiments for various shapes of buildings. The parameter estimation methods, i.e. the method of moments and quantile plot, are also reviewed, and the useful aspects of each method are discussed. Additionally, a simple worksheet for parameter estimation is derived based on the method of moment for practical application, and the accuracy is discussed comparing with a set of previously proposed formulae.

• Wind and Structures
• /
• 제23권4호
• /
• pp.275-300
• /
• 2016

Synchronous multi-pressure measurements were carried out with relatively long time duration for a double-layer reticulated shell roof model in the atmospheric boundary layer wind tunnel. Since the long roof is open at two ends for the storage of coal piles, three different testing cases were considered as the empty roof without coal piles (Case A), half coal piles inside (Case B) and full coal piles inside (Case C). Based on the wind tunnel test results, non-Gaussian time-dependent statistics of net wind pressure on the shell roof were quantified in terms of skewness and kurtosis. It was found that the direct statistical estimation of high-order moments and peak factors is quite sensitive to the duration of wind pressure time-history data. The maximum value of COVs (Coefficients of variations) of high-order moments is up to 1.05 for several measured pressure processes. The Mixture distribution models are proposed for better modeling the distribution of a parent pressure process. With the aid of mixture parent distribution models, the existing translated-peak-process (TPP) method has been revised and improved in the estimation of non-Gaussian peak factors. Finally, non-Gaussian peak factors of wind pressure, particularly for those observed hardening pressure process, were calculated by employing various state-of-the-art methods and compared to the direct statistical analysis of the measured long-duration wind pressure data. The estimated non-Gaussian peak factors for a hardening pressure process at the leading edge of the roof were varying from 3.6229, 3.3693 to 3.3416 corresponding to three different cases of A, B and C.

• 한국음향학회지
• /
• 제16권6호
• /
• pp.25-35
• /
• 1997

음성 신호 처리에 널리사용되어 온 2차 통계에 의한 음성 분석 방법은 잡음 환경에서 성능이 크게 저하되는 단점을 지닌다. 이에 반하여 고차 통계 방법은 Gaussian 잡음 등을 억제하는 특성을 가지고 있어서 잡음 환경에 상대적으로 강인한 음성 특징 추출을 가능하게 한다. 본 논문에서는 고차 통계에 의한 음성 분석 방법을 이용하여 백색 및 유색 잡음 환경에서의 문맥 독립형(text-independent) 화자식별 시스템을 제안하고, 기존의 2차 통계에 의한 방식과 성능을 비교하였다. 본 논문에서의 화자식별 시스템은 벡터 양자화 방법에 기반을 두고 있으며, 고차 통계 방법에 의한 유성음/무성음 판별을 통해 non-Gaussian 특징을 가지면서도 화자 정보가 집중되어 있는 유성음 부분에 대해서만 음성 특징을 추출하여 인식에 사용하였다. 50명의 화자를 대상으로 한 화자식별 실험 결과, 고차 통계 방법이 2차 통계에 의한 방법보다 잡음 환경에서 상대적으로 우수한 인식 성능을 나타냄을 확인하였다.

• 소음진동
• /
• 제8권4호
• /
• pp.715-722
• /
• 1998

An investigation into the stochastic bifurcation and response statistics of an autoparameteric system under broad-band random excitation is made. The specific system examined is a spring-pendulum system with internal resonance, which is known to be a good model for a variety of engineering systems, including ship motions with nonlinear coupling between pitching and rolling motions. The Fokker-Planck equations is used to genrage a general first-order differential equation in the dynamic moment of response coordinates. By means of the Gaussian and non-Gaussian closure methods the dynamic moment equations for the random responses of the system are reduced to a system of autonomous ordinanary differential equations. In view of equilibrium solutions of this system and their stability we examine the stochastic bifurcation and response statistics. The analytical results are compared with results obtained by Monte Carlo simulation.

• Journal of the Korean Statistical Society
• /
• 제25권4호
• /
• pp.515-527
• /
• 1996

It has been known for mean field models that the limiting distribution reflecting the asymptotic behavior of the system is non-Gaussian at the critical state. Recently, however, Papangelow showed for the critical Curie-Weiss mean field model that there exist Gaussian structures in the asymptotic behavior of the total magnetization. We construct Gaussian structures existing in the internal fluctuation of the system for the critical case of a generalized Curie-Weiss mean field model.

• 한국산업융합학회 논문집
• /
• 제3권1호
• /
• pp.43-51
• /
• 2000

The response statistics of a hinged-clamped beam under broad-band random excitation is investigated. The random excitation is applied at the nodal point of the second mode. By using Galerkin's method the governing equation is reduced to a system of nonautonomous nonlinear ordinary differential equations. A method based upon the Markov vector approach is used to generate a general first-order differential equation in the dynamic moment of response coordinates. By means of the Gaussian and non-Gaussian closure methods the dynamic moment equations for the random responses of the system are reduced to a system of autonomous ordinary differential equations. The case of two mode interaction is considered in order to compare it with the case of three mode interaction. The analytical results for two and three mode interactions are also compared with results obtained by Monte Carlo simulation.