• Journal of the Korean Data and Information Science Society
• /
• v.24 no.1
• /
• pp.179-187
• /
• 2013

Bates and Watts (1981) have discussed the problems of reparameterizing nonlinear models in obtaining accurate linear approximation confidence regions for the parameters. A similar problem exists with computing confidence curves for fitted values or predictions. The statistical behavior of fitted values does not depend on the parameterization. Thus, as long as the intrinsic curvature is small, standard Wald intervals for fitted values are likely to be sufficient. Accuracy of linear approximation for fitted values is investigated using confidence curves.

• Journal of the Korea institute for structural maintenance and inspection
• /
• v.6 no.1
• /
• pp.255-262
• /
• 2002

This paper presents a statistical analysis of empirical fragility curves for bridge. The empirical fragility curves are developed utilizing bridge damage data obtained from the 1995 Hyogoken Nanbu(Kobe) earthquake. Two-parameter lognormal distribution functions are used to represent the fragility curves with the parameters estimated by the maximum likelihood method. This paper also presents methods of testing the goodness of fit of the fragility curves and estimating the confidence intervals of the two parameters(median and log-standard deviation) of the distribution. An analytical interpretation of randomness and uncertainty associated with the median is provided.

• Journal of the Korean Statistical Society
• /
• v.32 no.1
• /
• pp.1-10
• /
• 2003

We consider obtaining graphical summaries of uncertainty in estimates of parameters in nonlinear models. A nonlinear constrained optimization algorithm is developed for likelihood based confidence intervals for the functions of parameters in the model The results are applied to the problem of finding significance levels in nonlinear models.

• Communications for Statistical Applications and Methods
• /
• v.16 no.1
• /
• pp.163-168
• /
• 2009

Given the specific mean shift outlier model, the standard approaches to obtaining test statistics for outliers are discussed. Accuracy of outlier tests is investigated using subset curvatures. These subset curvatures appear to be reliable indicators of the adequacy of the linearization based test. Also, we consider obtaining graphical summaries of uncertainty in estimating parameters through confidence curves. The results are applied to the problem of assessing the accuracy of outlier tests.

• Proceedings of the Korea Concrete Institute Conference
• /
• 2008.11a
• /
• pp.821-824
• /
• 2008

This paper presents a systematic approach for estimating material performance of concrete mixture design based on satisfaction curves developed from statistical evaluation of existing or newly obtained material property related data. In performance based material design (PBMD) method, concrete material used for construction of a structure is designed considering a structure's specified performance requirements based on its usage and characteristics such as environmental conditions, structure types, expected design life, etc.Satisfaction curves express the probabilities that one component of substrates (i.e., aggregate size, cement content, etc) of concrete mixture will sustain different criterion value for a given concrete mixture design. This study presents a statistical analysis method for setting up concrete material parameter versus concrete criterion relationships in the form of satisfaction curves and for estimating confidence bounds on these satisfaction curves. This paper also presents an analysis method to combine multiple satisfaction curves to form one unique satisfaction curve that can relate the performance of concrete to a single evaluating value. Based on several evaluated mixture design examples for various material properties, the validity of the proposed method is discussed in detail.

• IE interfaces
• /
• v.16 no.1
• /
• pp.103-110
• /
• 2003

Growth curves are widely used in forecasting the market demand. When there are only a few data points available, the estimated model parameters have a low confidence. In this case, if some expert opinions are available, it would be better for predicting future demand to adjust the model parameters using these information. This paper proposes the methodology for re-estimation of model parameters in growth curves when adjusting market potential and/or time of maximum sales. We also provide the detailed procedures for five growth curves including Bass, Logistic, Gompertz, Weibull and Cumulative Lognormal models. Applications to real data are also included.

• Journal of The Korean Astronomical Society
• /
• v.29 no.spc1
• /
• pp.217-218
• /
• 1996

The results of the Crab pulsar observations with the photometrical MANIA (Multichannel Analysis of Nanosecond Intensity Alterations) complex at the 6-m telescope are presented. More than 12 millions photons in UBVR-bands simultaneously with time resolution of $10^{-7}s$ were detected. Using the original software for search for optical pulsar period, we obtained the light curves of the object with time resolution of about 3.3 ${\mu}s$. Their detailed analysis gives the spectral change during pulse and subpulse, the shape of the pulse peaks, which are plateaus (with the duration of about 50${\mu}s$ for the main pulse), limits for an amplitude of fine temporal (stochastic and regular) structure of pulse and sub pulse and the interpulse space intensity. The results of CCD-spectroscopy of the Crab pulsar show that its summarized spectrum is flat. There are no lines, neither emission nor absorbtion ones. Upper limit for line intensity or depth is $3.5\%$ with the confidence probability of $95\%$.

• Transactions of the Korean Society of Automotive Engineers
• /
• v.12 no.2
• /
• pp.117-122
• /
• 2004

The fatigue life in wheels was predicted by simulating the experimental method using Finite-Element analysis. Based on a high frequency fatigue property, calculations of the stresses in wheels were performed by simulating the rotating bending fatigue test. Wheels made of an aluminum alloy(A356.2) were tested using a bending fatigue tester. Results from bending fatigue test showed a linear correlation between bending moment and stress amplitude. Consequently, Finite-Element calculations were performed by a linear analysis. In order to find stress-cycles curves, spoke parts of wheel were tested using a rotary bending fatigue tester. Also, highly accurate Finite-Element analysis requires regression lines and confidence intervals from these results. In conclusion, if the fatigue data related to the material and manufacturing procedure are reliable, the prediction on fatigue lift in wheels can be carried out with high accuracy.

• Communications for Statistical Applications and Methods
• /
• v.5 no.2
• /
• pp.491-501
• /
• 1998

Motivated by bioequivalence studies which involve comparisons of pharmaceutically equivalent dosage forms, we propose a more general decision rule for showing equivalence simultaneously between multiple means and a control mean. Namely, this testing procedure is concerned with the situation in that one must make decisions as to the bioequivalence of an original drug product and several generic formulations of that drug. This general test is developed by considering a spherical confidence region, which is a direct extension of the usual t-based confidence interval rule formally approved by the U.S. Food and Drug Administration. We characterize the test by the probability of rejection curves and assess its performance via Monte-Carlo simulation. Since the manufacturer's main concern is the proper choice of sample sizes, we provide optimal sample sizes from the Monte-Carlo simulation results. We also consider an application of the generalized equivalence test to a repeated measures design.

• Computers and Concrete
• /
• v.7 no.5
• /
• pp.455-468
• /
• 2010

This paper presents a prediction and simulation method of tensile stress-strain curves of Engineered Cementitious Composites (ECC). For this purpose, the bridging stress and crack opening relations were obtained by the fiber bridging constitutive law which is quantitatively able to consider the fiber distribution characteristics. And then, a multi-linear model is employed for a simplification of the bridging stress and crack opening relation. In addition, to account the variability of material properties, randomly distributed properties drawn from a normal distribution with 95% confidence are assigned to each element which is determined on the basis of crack spacing. To consider the variation of crack spacing, randomly distributed crack spacing is drawn from the probability density function of fiber inclined angle calculated based on sectional image analysis. An equation for calculation of the crack spacing that takes into quantitative consideration the dimensions and fiber distribution was also derived. Subsequently, a series of simulations of ECC tensile stress-strain curves was performed. The simulation results exhibit obvious strain hardening behavior associated with multiple cracking, which correspond well with test results.