• Journal of Multimedia Information System
• /
• 제8권1호
• /
• pp.31-34
• /
• 2021

The classical HMM is defined by a parameter triple �� = (��, A, B), where each parameter represents a collection of probability distributions: initial state, state transition and output distributions in order. This paper proposes a new stationary parameter e = (e1, e2, …, eN) where N is the number of states and et = P(|xt = i, y) for describing how an input pattern y ends in state xt = i at time t followed by nothing. It is often said that all is well that ends well. We argue here that all should end well. The paper sets the framework for the theory and presents an efficient inference and training algorithms based on dynamic programming and expectation-maximization. The proposed model is applicable to analyzing any sequential data with two or more finite segmental patterns are concatenated, each forming a context to its neighbors. Experiments on online Hangul handwriting characters have proven the effect of the proposed augmentation in terms of highly intuitive segmentation as well as recognition performance and 13.2% error rate reduction.

• Geomechanics and Engineering
• /
• 제37권6호
• /
• pp.591-597
• /
• 2024

The purpose of this paper is to depict the effect of gravity on a nonlocal thermoelastic medium with initial stress. The Lord-Shulman and Green-Lindsay theories with fractional derivative order serve as the foundation for the formulation of the fundamental equations for the problem. To address the problem and acquire the exact expressions of physical fields, appropriate non-dimensional variables and normal mode analysis are used. MATLAB software is used for numerical calculations. The projected outcomes in the presence and absence of the gravitational field, along with a nonlocal parameter, are compared. Additional comparisons are made for various fractional derivative order values. It is evident that the variation of physical variables is significantly influenced by the fractional derivative order, nonlocal parameter and gravity field.

• 한국통계학회:학술대회논문집
• /
• 한국통계학회 2002년도 춘계 학술발표회 논문집
• /
• pp.25-28
• /
• 2002

비선형모형분석의 초기 단계에서 초기값(starting value, initial parameter value)를 결정하는 문제는 비선형모형의 모수추정을 위한 반복기법의 수렴속도나 국소값(local minimum)문제에 영향을 주게 된다. 본 논문을 통하여 탐색적 자료분석이 초기값를 결정하는 데 도움을 줄 수 있음을 보이고자 한다.

• Communications for Statistical Applications and Methods
• /
• 제6권3호
• /
• pp.667-676
• /
• 1999

We investigate the effects of omission of initial observations in each individuals in the panel data regression model when the disturbances follow a serially correlated one way error components. We show that the first transformed observation can have a relative large hat matrix diagonal component and a large influence on parameter estimates when the correlation coefficient is large in absolute value.

• 대한전자공학회논문지SP
• /
• 제47권1호
• /
• pp.112-119
• /
• 2010

본 논문에서는 스테레오 이미지로부터 디스패리티 맵을 추출하기 위한 확률모델을 제시하고 이의 해를 구하는 과정은 에너지 기반 스테레오 정합과 일치함을 이론적으로 증명한다. 정합되는 화소 간의 차와 인근 화소에 해당되는 디스패리티의 차는 exponential 확률분포에 근사하다는 사실을 실험적으로 확인하고 이에 근거하여 이들의 정합 파라미터를 최적화하는 식을 유도하고 이를 실험적으로 구하는 방법을 제시한다. 에너지 기반 스테레오 정합 알고리즘의 성능은 기본적으로 정합 파라미터의 크기에 매우 민감하므로 이미지에 따라 적절한 값을 사전에 구하여 적용하여야 한다. 제안한 방식은 초기에 임의의 파라미터로 디스패리티 맵을 구한 후에 이의 통계적 특성을 이용하여 정합 파라미터를 추정하고 추정된 파라미터를 적용하여 디스패리티 맵을 재차 구하는 과정을 반복함으로써 최적의 파라미터에 적응적으로 수렴하도록 조정한다. 따라서, 이미지에 따라 사전에 정합 파라미터를 구하여야 하는 문제를 해결할 수 있다. Middlebury 웹사이트에서 제공한 다양한 스테레오 이미지를 이용하여 제안한 방식으로 구한 파라미터가 최적의 값으로 수렴하는지를 조사하고 이의 수렴 속도와 성능 개선 효과 등을 확인한다.

• 소성∙가공
• /
• 제20권6호
• /
• pp.450-455
• /
• 2011

This paper presents an inverse design methodology for the cross section geometry of mold spring with a rectangular cross section as the starting material for a coiling process. The cross-sections of mold springs are universally rectangular, as the parallel sides minimize the possibility of failure under high service loads. Pre-coiled wires are initially designed to have a trapezoidal cross section, which becomes a rectangle by the coiling process. This study demonstrates a numerical exercise to predict changes in the sectional geometry in spring manufacture and to obtain the initial cross section which becomes the exact rectangle desired from the manufacturing process. Finite element analysis was carried out to calculate the sectional changes for various mold springs. Geometrical parameters were the widths at inner and outer radii, the inner and the outer corner radii, and the height. A partial least square regression analysis was carried out to find the main contributing factors for deciding initial design values. The height and the width mainly affected various initial parameters. The initial width at the inner radius was mostly affected by various specification parameters.

• 대한기계학회논문집B
• /
• 제29권1호
• /
• pp.55-62
• /
• 2005

Combustion Characteristics and quenching criteria of micro combustor in various condition was exploited experimentally. Two different gases were used, and various geometric matrixes were considered to figure out quenching characteristic of micro combustor. The micro combustor studied in this study was constant volume, and has cylindrical shape. Geometric parameter of combustor was defined to be combustor height and diameter. The effect of height was exploited parametrically as 1 mm, 2mm and 3mm and the effect of diameter was parameterized to be 7.5mm and 15mm. Three different combustibles. (1) Stoichiometric mixture of methane and are, (2) Stoichiometric mixture of hydrogen and air and (3) Mixture of hydrogen and air with fuel stoichiometry of two were used. Pressure transition during combustion process was recovered. The ratio of maximum pressure to initial pressure responded favorably with the change of height of combustor and the initial pressure, the maximum pressure was also increased. The flame propagation was observed only when a specific condition was satisfied. From the experiment the condition that guarantees stable propagation of flame was tabulated. The tabulated results includes criteria of quenching according to combustor height, combustor diameter, species of fuel and initial pressure.

• Structural Engineering and Mechanics
• /
• 제29권5호
• /
• pp.469-488
• /
• 2008

Finite element methods have often been used for structural analyses of various mechanical problems. When finite element analyses are utilized to resolve mechanical systems, numerical uncertainties in the initial data such as structural parameters and loading conditions may result in uncertainties in the structural responses. Therefore the initial data have to be as accurate as possible in order to obtain reliable structural analysis results. The typical finite element method may not properly represent discrete systems when using uncertain data, since all input data of material properties and applied loads are defined by nominal values. An interval finite element analysis, which uses the interval arithmetic as introduced by Moore (1966) is proposed as a non-stochastic method in this study and serves a new numerical tool for evaluating the uncertainties of the initial data in structural analyses. According to this method, the element stiffness matrix includes interval terms of the lower and upper bounds of the structural parameters, and interval change functions are devised. Numerical uncertainties in the initial data are described as a tolerance error and tree graphs of uncertain data are constructed by numerical uncertainty combinations of each parameter. The structural responses calculated by all uncertainty cases can be easily estimated so that structural safety can be included in the design. Numerical applications of truss and frame structures demonstrate the efficiency of the present method with respect to numerical analyses of structural uncertainties.

• Structural Engineering and Mechanics
• /
• 제90권4호
• /
• pp.357-370
• /
• 2024

Due to the fact that the mechanism of the effects of temperature and initial geometric imperfection on low-velocity impact problem of axially moving plates is not yet clear, the present paper is to fill the gap. In the present paper, the nonlinear dynamic behavior of axially moving imperfect graphene platelet reinforced metal foams (GPLRMF) plates subjected to lowvelocity impact in thermal environment is analyzed. The equivalent physical parameters of GPLRMF plates are estimated based on the Halpin-Tsai equation and the mixing rule. Combining Kirchhoff plate theory and the modified nonlinear Hertz contact theory, the nonlinear governing equations of GPLRMF plates are derived. Under the condition of simply supported boundary, the nonlinear control equation is discretized with the help of Gallekin method. The correctness of the proposed model is verified by comparison with the existing results. Finally, the time history curves of contact force and transverse center displacement are obtained by using the fourth order Runge-Kutta method. Through detailed parameter research, the effects of graphene platelet (GPL) distribution mode, foam distribution mode, GPL weight fraction, foam coefficient, axial moving speed, prestressing force, temperature changes, damping coefficient, initial geometric defect, radius and initial velocity of the impactor on the nonlinear impact problem are explored. The results indicate that temperature changes and initial geometric imperfections have significant impacts.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2003년도 ICCAS
• /
• pp.752-757
• /
• 2003

Recently, 3D structure recovery through self-calibration of camera has been actively researched. Traditional calibration algorithm requires known 3D coordinates of the control points while self-calibration only requires the corresponding points of images, thus it has more flexibility in real application. In general, self-calibration algorithm results in the nonlinear optimization problem using constraints from the intrinsic parameters of the camera. Thus, it requires initial value for the nonlinear minimization. Traditional approaches get the initial values assuming they have the same intrinsic parameters while they are dealing with the situation where the intrinsic parameters of the camera may change. In this paper, we propose new initialization method using the minimum 2 images. Proposed method is based on the assumption that the least violation of the camera’s intrinsic parameter gives more stable initial value. Synthetic and real experiment shows this result.