• Journal of the Korean Statistical Society
• /
• v.36 no.2
• /
• pp.321-333
• /
• 2007

In the proportional hazard model with the beta process prior, the posterior computation with the discrete approximation is considered. The time period of interest is partitioned by small intervals. On each partitioning interval, the likelihood is approximated by that of a binomial experiment and the beta process prior is by a beta distribution. Consequently, the posterior is approximated by that of many independent binomial model with beta priors. The analysis of the leukemia remission data is given as an example. It is illustrated that the length of the partitioning interval affects the posterior and one needs to be careful in choosing it.

• Proceedings of the Korean Society for Bioinformatics Conference
• /
• 2005.09a
• /
• pp.357-360
• /
• 2005

In this paper we consider the well-known semiparametric proportional hazards (PH) models for survival analysis. These models are usually used with few covariates and many observations (subjects). But, for a typical setting of gene expression data from DNA microarray, we need to consider the case where the number of covariates p exceeds the number of samples n. For a given vector of response values which are times to event (death or censored times) and p gene expressions (covariates), we address the issue of how to reduce the dimension by selecting the significant genes. This approach enable us to estimate the survival curve when n < < p. In our approach, rather than fixing the number of selected genes, we will assign a prior distribution to this number. The approach creates additional flexibility by allowing the imposition of constraints, such as bounding the dimension via a prior, which in effect works as a penalty. To implement our methodology, we use a Markov Chain Monte Carlo (MCMC) method. We demonstrate the use of the methodology to diffuse large B-cell lymphoma (DLBCL) complementary DNA(cDNA) data.

• Journal of the Korean Operations Research and Management Science Society
• /
• v.21 no.1
• /
• pp.71-80
• /
• 1996

This paper presents a stochastic partial inventory model for the situation in which demand is deterministic, lead time follows normal distribution and backorder ratio during the stockout period decreases exponentially according to the length of backorder period. In this situation, an objective function is formulated to minimize the average annual cost, which is the sum of the ordering, carrying time-proportional backordering, quantity-proportional backordering and lost sales costs. And then the procedure of iterative solution method for the model is developed to find optimal reorder point and order quantity and numerical example to illustrate the proposed method is presented.

• Communications for Statistical Applications and Methods
• /
• v.17 no.1
• /
• pp.117-126
• /
• 2010

Clinical researchers often analyze survival data as binary outcomes using the logistic regression method. This paper examines the information loss resulting from analyzing survival time as binary outcomes. We first demonstrate that, under the proportional hazard assumption, this binary discretization does result in a significant information loss. Second, when fitting a logistic model to survival time data, researchers inadvertently use the maximal statistic. We implement a numerical study to examine the properties of the reference distribution for this statistic, finally, we show that the logistic regression method can still be a useful tool for analyzing survival data in particular when the proportional hazard assumption is questionable.

• Proceedings of the Korean Society for Technology of Plasticity Conference
• /
• 2004.10a
• /
• pp.230-233
• /
• 2004

The purpose of this study is to establish the optimal induction heating conditions of various preform types used for TR forging. The finite element model coupled electro-magnetic and transient heat transfer was employed to evaluate the distribution of temperature at the billet. Power control method was applied to control temperature of preform in induction heating because TR forging is not a continuous process. Power schedule that consists of heating and holding stage was suggested. In heating stage, power is inversely proportional to diameter of preform but the time of heating stage is directly proportional to the diameter of preform. But, in holding stage, the required power for thermal equilibrium per unit volume of the billet decreases with an increase in a diameter of billet due to the increase of efficiency.

• Korean Journal of Air-Conditioning and Refrigeration Engineering
• /
• v.22 no.7
• /
• pp.468-473
• /
• 2010

This paper deals with approximate integral solutions to the one-dimensional model describing the charging process of stratified thermal storage tanks. Temperature is assumed to be the form of Fermi-Dirac distribution function, which can be separated to two sets of cubic polynomials for each hot and cold side of thermal boundary layers. Proposed approximate integral solutions are compared to the previous works of the approximate analytic solutions and show reasonable agreement. The approach, however, has benefits in mathematical difficulties, complicated solution form and unstable convergence of series solution founded in the previous analytic solutions. Solutions for a semi-infinite region, which have simple closed form solutions, give close agreement to those for a finite region. Thermocline thickness is obtained in closed form and shows proportional behavior to the square root of time and inverse proportional behavior to the square root of flow rate.

• Journal of the Korean Society of Industry Convergence
• /
• v.4 no.2
• /
• pp.157-166
• /
• 2001

The purpose of this study is to examine the accuracy of the code provisions on lateral load distribution factors of prestressed concrete girder bridges. Most designers in Korea use the lever method or lateral load distribution formula in the existing design codes. However, the methods do not account for the effect of bridge skew or direction of diaphragm. Therefore, this study analysed the prestressed concrete girder bridge with grillage model for various girder spacings, directions of diaphragms, span lengths, and skews, and compared the results with those of existing design code. It has been found that lateral load distribution factors were proportional to the girder spacing while they were not significantly affected by the change of span length, direction of diaphragm, and skew. For bending moments, lateral load distribution factors from the grillage analysis were 60%~68% of those from Korean bridge design code. Therefore, the code provisions result in very conservative design. For support reactions, however, lateral load distribution factors from the grillage analysis were slightly greater than those from Korean bridge design code. Therefore, the capacity of bearings of the bridge with a large skew should be determined by grillage analysis.

• Korea-Australia Rheology Journal
• /
• v.15 no.1
• /
• pp.35-41
• /
• 2003

It is expected that particle size distribution of any portion obtained through screening, is of more uniform than that of the original mixture, typically following such as log-normal, Rosin-Rammler distributions and so on. In this study, therefore, a new relation between parameters of the uniform distribution and flow characteristics of the coarse particle suspensions is derived based on the continuous polydisperse model (Ookawara and Ogawa, 2002b), which is derived from the discrete polydisperse model (Ookawara and Ogawa,2002a). The derived model equation predicts a linear increase of viscosity with shear rate, viz., dilutant flow characteristics. Further, the increase of viscosity is expected to be proportional to the square of volume fraction of particles, and to show the linear dependency on density and average diameter of particles. It is also shown that the uniform distribution model includes additional term that expresses the effect of distribution width. For verification of the model, the experimental results of Clarke (1967) are cited as well as in our previous work for the monodisperse model (Ookawara and Ogawa,2000) since most parameters were varied independently in his work. It is suggested that the newly introduced term expands the applicable range compared with the monodisperse model.

• Journal of Korean Society of Forest Science
• /
• v.98 no.5
• /
• pp.576-583
• /
• 2009

Based on the stem analysis and biomass measurement of 36 trees and 1,576 branches in Pinus sylvestris var. mongolica (Mongolian pine) plantations of Northeast China, this study was conducted to develop estimation model equation for leaf biomass of a single tree and branch, to examine the vertical distribution of leaf biomass in the crown, and to evaluate the proportional ratios of biomass by tree parts, stem, branch, and leaf. The results indicated that DBH and crown length were quite appropriate to estimate leaf biomass. The biomass of single branch was highly correlated with branch collar diameter and relative height of branch in the crown, but not much with stand density, site quality, and tree height. Weibull distribution function would have been appropriate to express vertical distribution of leaf biomass. The shape parameters from 29 sample trees out of 36 were less than 3.6, indicating that vertical distribution of leaf biomass in the crown was displayed by bell-shaped curve, a little inclined toward positive side. Apparent correlationship was obtained between leaf biomass and branch biomass having resulted in linear function equation. The stem biomass occupied around 80% and branch and leaf made up about 20% of total biomass in a single tree. As the level of tree class was increased from class I to class V, the proportion of the stem biomass to total biomass was gradually increased, but that of branch and leaf became decreased.

• Journal of Mechanical Science and Technology
• /
• v.16 no.9
• /
• pp.1084-1092
• /
• 2002

The frequency dependency of Rayleigh surface wave is investigated indirectly by measuring the angular dependency of the backward radiation of the incident ultrasonic wave in two kinds of degraded specimens by scuffing or corrosion. Then, the frequency dependency is compared with the residual stress distribution or the corrosion-fatigue characteristics for the scuffed or corroded specimens, respectively. The width of the backward radiation profile increases with the increase of the variation in residual stress distribution for the scuffed specimens. In the corroded specimens, the profile width decreases with the increase of the effective aging layer thickness and is inversely proportional to the exponent, m, in the Paris' law that can predict the crack size increase due to fatigue. The result observed in this study demonstrates high potential of backward radiated ultrasound as a tool for nondestructive evaluation of subsurface gradient of material degradation generated by scuffing or corrosion.