• Communications for Statistical Applications and Methods
• /
• v.12 no.3
• /
• pp.729-742
• /
• 2005

In this paper, we discuss the propriety of the various noninformative priors for the Pareto distribution. The reference prior, Jeffreys prior and ad hoc noninformative prior which is used in several literatures will be introduced and showed that which prior gives the proper posterior distribution. The reference prior and Jeffreys prior give a proper posterior distribution, but ad hoc noninformative prior which is proportional to reciprocal of the parameters does not give a proper posterior. To compute survival function, we use the well-known approximation method proposed by Lindley (1980) and Tireney and Kadane (1986). And two methods are compared by simulation. A real data example is given to illustrate our methodology.

• Journal of the Korean Society of Manufacturing Technology Engineers
• /
• v.22 no.2
• /
• pp.241-247
• /
• 2013

The doctor blade is equipped in a toner cartridge and is a device to maintain the uniform thickness of a toner by controlling the pressure on the developing roller. The contact pressure between the developing roller and the doctor blade is one of the significant factors for image quality and durability of toner cartridge. The purpose of this study is to develop approximation model in order to minimize the time and cost which are needed much required in making optimal design of the doctor blade. Central composite design was used for the design of experiment and response surface design was used for approximation. The data for contact pressure were acquired through finite element analysis and data of image density and toner weight were acquired through experiment. The approximation model developed in this study has presented very high fitness.

• Journal of Mechanical Science and Technology
• /
• v.15 no.8
• /
• pp.1132-1142
• /
• 2001

The paper describes the construction of global function approximation models for use in design optimization via global search techniques such as genetic algorithms. Two different approximation methods referred to as evolutionary fuzzy modeling (EFM) and neuro-fuzzy modeling (NFM) are implemented in the context of global approximate optimization. EFM and NFM are based on soft computing paradigms utilizing fuzzy systems, neural networks and evolutionary computing techniques. Such approximation methods may have their promising characteristics in a case where the training data is not sufficiently provided or uncertain information may be included in design process. Fuzzy inference system is the central system for of identifying the input/output relationship in both methods. The paper introduces the general procedures including fuzzy rule generation, membership function selection and inference process for EFM and NFM, and presents their generalization capabilities in terms of a number of fuzzy rules and training data with application to a three-bar truss optimization.

• Journal of applied mathematics & informatics
• /
• v.12 no.1_2
• /
• pp.155-164
• /
• 2003

A method is proposed to predict the distances between given residue pairs (between C$\sub$${\alpha}$/ atoms) of a protein using a sequence to structure mapping by indefinite quadratic approximation. The prediction technique requires a data fitting in three dimensional space with coordinates of the residues of known structured proteins and leads to a numerical ref resentation of 20 amino acids by minimizing a large least norm iteratively. These approximations are used in distance prediction for given residue pairs. Some computational experience on a test set of small proteins from Brookhaven Protein Data Bank are given.

• Proceedings of the Computational Structural Engineering Institute Conference
• /
• 2001.10a
• /
• pp.93-100
• /
• 2001

The three dimensional(3D) reconstruction from two dimensional(2D) image data is using in many fields such as RPD(Rapid Product Development) and reverse engineering. In this paper, the main step of 3D reconstruction is comprised of two steps : image processing step and B-spline surface approximation step. In the image processing step, feature points of each cross-section are obtained by means of several image processing technologies. In the B-spline surface approximation step, using the data of feature points obtained in the image processing step, the control points of B-spline surface are obtained, which are used for IGES file of 3D CAD model.

• The Transactions of the Korea Information Processing Society
• /
• v.4 no.3
• /
• pp.747-753
• /
• 1997

Function approximation from a set of input-output parirs has numerous applications in scientiffc and engineer-ing areas.Multiayer feedforward neural networks have been proposed as a good approximator of noninear function.The back propagation (BP) algorithm allows muktiayer feedforward neural networks oro learn input-output mappongs from training samples.However, the mapping acquired through the BP algorithm nay be cor-rupt when errorneous trauning data are employed.In this paper we propose a robust BP learning algorithm that is resistant to the errormeous data and is capable of rejecting gross errors during the approximation process.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.34 no.2
• /
• pp.30-34
• /
• 2011

Variable precision rough set models have been successfully applied to problems whose domains are discrete values. However, there are many situations where discrete data is not available. When it comes to the problems with interval values, no variable precision rough set model has been proposed. In this paper, we propose a variable precision rough set model for interval values in which classification errors are allowed in determining if two intervals are same. To build the model, we define equivalence class, upper approximation, lower approximation, and boundary region. Then, we check if each of 11 characteristics on approximation that works in Pawlak's rough set model is valid for the proposed model or not.

• Journal of the Korean Data and Information Science Society
• /
• v.17 no.3
• /
• pp.753-762
• /
• 2006

The Snakes and GVF used to find object edges dynamically have assigned their initial contour arbitrarily. If the initial contours are located in the neighboring regions of object edges, Snakes and GVF can be close to the true boundary. If not, these will likely to converge to the wrong result. Therefore, this paper proposes a new initialization of Snakes and GVF using convex hull approximation, which initializes the vertex of Snakes and GVF as a convex polygonal contour near object edges. In simulation result, we show that the proposed algorithm has a faster convergence to object edges than the existing methods. Our algorithm also has the advantage of extracting whole edges in real images.

• Proceedings of the KSRS Conference
• /
• 2005.10a
• /
• pp.276-279
• /
• 2005

As the Ubiquitous generation approaches, the importance of the sensor data processing is growing. The data approximation scheme, one of the data processing methods, can be the key of sensor data processing, for it is related not only to the lifetime of sensors but also to the size of the storage. In this paper, we propose the Harmonic Wavelet transform which can minimize the relative error for given sensor data. Harmonic Wavelets use the harmonic mean as a representative which is the minimum point of the maximum relative error between two data values. In addition, Harmonic Wavelets retain the relative errors as wavelet coefficients so we can select proper wavelet coefficients that reduce the relative error more easily. We also adapt the greedy algorithm for local optimization to reduce the time complexity. Experimental results show the performance and the scalability of Harmonic Wavelets for sensor data.

• The Journal of the Korea Contents Association
• /
• v.4 no.3
• /
• pp.81-90
• /
• 2004

Recently, It has been increased to use a multi-dimensional data in various applications with a rapid growth of the computing environment. In this paper, we propose the vector approximate tree for content-based retrieval of multi-dimensional data. The proposed index structure reduces the depth of tree by storing the many region information in a node because of representing region information using space partition based method and vector approximation method. Also it efficiently handles 'dimensionality curse' that causes a problem of multi-dimensional index structure by assigning the multi-dimensional data space to dynamic bit. And it provides the more correct regions by representing the child region information as the parent region information relatively. We show that our index structure outperforms the existing index structure by various experimental evaluations.