• Journal of the Korean Data and Information Science Society
• /
• v.10 no.1
• /
• pp.103-110
• /
• 1999

In this paper, we summarize some relative risk estimators under the two sample model with proportional hazard and examine the relative efficiencies of the nonparametric estimators relative to the maximum likelihood estimator of a parametric survival function under random censoring model by comparing their asymptotic variances.

• Journal of the Korean Data and Information Science Society
• /
• v.13 no.1
• /
• pp.47-53
• /
• 2002

A new redescending M-estimating function is introduced. The estimators by this new redescending function attain the same level of robustness as the existing redescending M-estimators, but have less asymptotic variances than others except few cases. We have focused on estimating a location parameter, but the method can be extended for a scale estimation.

• Journal of the Korean Data and Information Science Society
• /
• v.9 no.2
• /
• pp.203-209
• /
• 1998

By assuming a progressively censored sample, we propose the approximate maximum likelihood estimator (AMLE) of the location nd the scale parameters of the two-parameter normal distribution and obtain the asymptotic variances and covariance of the AMLEs. An example is given to illustrate the methods of estimation discussed in this paper.

• Journal of the Korean Statistical Society
• /
• v.30 no.1
• /
• pp.99-113
• /
• 2001

This paper deals with the problem of estimating the means in two inverse Gaussian populations with equal but unknown coefficient of variation. The maximum likelihood estimators are derived by solving a cubic equation and their asymptotic variances are presented for comparative purpose. Monte-Carlo simulation is conducted to investigate the efficiency of the estimators relative to the sample means over a wide range of values for the sample size and the coefficient of variation. The effect on this efficiency under the departure from the assumption of common coefficient of variation is also studied.

• Journal of the military operations research society of Korea
• /
• v.22 no.2
• /
• pp.215-226
• /
• 1996

The estimation of lifetimes are examined when the distribution of lifetimes are Weibull. It is assumed that, due to physical restrictions and/or economic requirements, the lifetimes are investigated only at certain time intervals during the test period with 'mixed replacement' experiment, even though it is well known that 'with replacement' experiment produces better accuracy than 'without replacement' one. The maximum likelihood estimators are derived through the iterative method like as Lawless(1982). Also Cramer-Rao lower bounds are found as the asymptotic variances of the estimates.

• Journal of Korean Society for Quality Management
• /
• v.18 no.1
• /
• pp.9-20
• /
• 1990

The purpose of this paper is to propose the modified semiparametric estimators for survival function in the Cox's regression model with randomly censored data based on Tsiatis and Breslow estimators, and present their asymptotic variances estimates. The proposed estimators are compared to Tsiatis, Breslow, and Kaplan-Meier estimators through a small-sample Monte Carlo study. The simulation results show that the proposed estimators are preferred for small sample sizes.

• Journal of the Korean Data and Information Science Society
• /
• v.10 no.2
• /
• pp.405-413
• /
• 1999

By assuming a type-II censoring, we propose the approximate maximum likelihood estimators (AMLEs) of the location and the scale parameters of the two-parameter Rayleigh distribution and calculate the asymptotic variances and covariance of the AMLEs.

• Water Engineering Research
• /
• v.1 no.1
• /
• pp.11-23
• /
• 2000

Regional flood frequency analysis has been developed by employing the nearby site's information to improve a precision in estimating flood quantiles at the site of interest. In this paper, single site and regional flood frequency analyses were compared based of the 2-parameter Weibull model. For regional analysis, two approaches were employed. The First one is to use the asymptotic variances of the quantile estimators derived based of the assumption that all sites including the site of interest are independent each other. This approach may give the maximum regional gain due to the spatial independence assumption among sites. The second one in Hosking's regional L-moment algorithm. These methods were applied to annual flood data. As the results, both methods generally showed the regional gain at the site of interest depending on grouping the sites as homogeneous. And asymptotic formula generally shows smaller variance than those from Hosking's algorithm. If the shape parameter of the site of interest from single site analysis is quite different from that from regional analysis then Hosking's results might be better than the asymptotic ones because the formula was derived based on the assumption that all sites have the same regional shape parameter. Furthermore, in such a case, regional analysis might be worse than single site analysis in the sense of precision of flood quantile estimation. Even though the selected sites may satisfy Hosking's criteria, regional analysis may not give a regional gain for specific and nonexceedance probabilities.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.28 no.6B
• /
• pp.723-729
• /
• 2008

Confidence intervals of growth curves are calculated to assess the uncertainty of index flood method as a regional frequency analysis. The asymptotic variance of quantile estimator for the generalized logistic distribution is introduced to evaluate confidence intervals. In addition, the variances of at-site frequency estimator and regional frequency estimator are used to evaluate an efficiency index. The efficiency indexes for 14 homogeneous regions based on 378 stations show that index flood method estimators are more efficient than at-site frequency estimators. It is shown that the number of sites in a region needs to be limited for regional gain.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.13 no.4
• /
• pp.151-161
• /
• 1993

The log-Gumbel distribution in real space is defined by transforming the conventional log-Gumbel distribution in log space. For this model, the parameter estimation techniques are applied based on the methods of moments, maximum likelihood and probability weighted moments. The asymptotic variances of estimator of the quantiles for each estimation method are derived to find the confidence limits for a given return period. Finally, the log-Gumbel model is applied to actual flood data to estimate the parameters, quantiles and confidence limits.