• Journal of the Korean Data and Information Science Society
• /
• v.20 no.5
• /
• pp.895-903
• /
• 2009

The penalized partial likelihood based on restricted maximum likelihood method has been widely used for the inference of frailty models. However, the standard-error estimate for frailty parameter estimator can be downwardly biased. In this paper we show that such underestimation can be corrected by using hierarchical likelihood. In particular, the hierarchical likelihood gives a statistically efficient procedure for various random-effect models including frailty models. The proposed method is illustrated via a numerical example and simulation study. The simulation results demonstrate that the corrected standard-error estimate largely improves such bias.

• Journal of the Korean Data and Information Science Society
• /
• v.17 no.3
• /
• pp.999-1007
• /
• 2006

The semiparametric gamma frailty models have been often used for multivariate survival analysis because they give an explicit marginal likelihood. The commonly used estimation procedure is the profile likelihood method based on marginal likelihood, which provides the same parameter estimates as the EM algorithm. In this paper we show in finite samples the standard profile-likelihood method can lead to an underestimation of parameters, particularly for the frailty parameter. To overcome this problem, we propose an adjusted profile-likelihood method. For the illustration a numerical example and a small-sample simulation study are presented.

• Journal of the Korean Data and Information Science Society
• /
• v.19 no.4
• /
• pp.1173-1181
• /
• 2008

In this paper, when the parameter of interest is the shape parameter in Pareto distribution, we develop likelihood based inference for this parameter. Specially, we develop signed log-likelihood ratio statistic and the modified signed log-likelihood ratio statistic for the shape parameter. It is well-known that as sample size grows, the modified signed log-likelihood ratio statistic converges to standard normal distribution faster than the signed log-likelihood ratio statistic. But the computation of the modified signed log-likelihood statistic is hard or even impossible when the sufficient statistics and the ancillary statistics are not clear. In this case, one can consider an approximation to the modified signed log-likelihood statistic. Specially, when the parameter of interest is informationally orthogonal to the nuisance parameters, we propose the approximate modified signed log-likelihood statistic. Through simulation, we investigate the performances of the proposed statistics with the signed log-likelihood statistic.

• Journal of the Korean Data and Information Science Society
• /
• v.24 no.4
• /
• pp.911-918
• /
• 2013

This study deals with the small sample likelihood based inference for the ratio of two normal variances. The small sample likelihood inference is an approximation method. The signed log-likelihood ratio statistic and the modified signed log-likelihood ratio statistic, which converge to standard normal distribution, are proposed for the normal variance ratio. Through the simulation study, the coverage probabilities of confidence interval and power of the exact, the signed log-likelihood and the modified signed log-likelihood ratio statistic will be compared. A real data example will be provided.

• Journal of the Korean Data and Information Science Society
• /
• v.9 no.2
• /
• pp.219-225
• /
• 1998

The parameters in linear models with censored normal responses are usually estimated by the iterative maximum likelihood and least square methods. However, the iterative least square method is simple but hardly has theoretical justification, and the iterative maximum likelihood estimating equations are complicatedly derived. In this paper, we justify these methods via Wedderburn (1974)'s quasi-likelihood approach. This provides an explicit justification for the iterative least square method and also directly the iterative maximum likelihood method for estimating the regression coefficients.

• Journal of the Korean Data and Information Science Society
• /
• v.8 no.2
• /
• pp.247-260
• /
• 1997

The quasi-likelihood models which greatly widened the scope of generalized linear models are widely used in data analysis where a likelihood is not available. Since a quasi-likelihood may not appear to be an ordinary likelihood for any known distribution in the natural exponential family, to fit the quasi-likelihood models the standard statistical packages such as GLIM, GENSTAT, S-PLUS and so on may not directly applied. SAS/IML is very useful for fitting of such models. In this paper, we present simple SAS/IML(version 6.11) program which helps to fit and analyze the quasi-likelihood models applied to the leaf-blotch data introduced by Wedderburn(1974), and the problem with deviance useful generally to model checking is pointed out, and then its solution method is mention through the data analysis based on this quasi-likelihood models checking.

• Communications for Statistical Applications and Methods
• /
• v.18 no.4
• /
• pp.485-493
• /
• 2011

The proportional likelihood ratio order is an extension of the likelihood ratio order for the non-negative absolutely continuous random variables. In addition, the Lindley distribution has been over looked as a mixture of two exponential distributions due to the popularity of the exponential distribution. In this paper, we first recalled the above concepts and then obtained various properties of the Lindley distribution due to the proportional likelihood ratio order. These results are more general than the likelihood ratio ordering aspects related to this distribution. Finally, we discussed the proportional likelihood ratio ordering in view of the weighted version of the Lindley distribution.

• Journal of the Korean Operations Research and Management Science Society
• /
• v.38 no.1
• /
• pp.201-216
• /
• 2013

Diffusion is a basic mathematical tool for modern financial engineering. The theory of the estimation methods for diffusion models is an important topic of the financial engineering. Many researches have been tried to apply the likelihood estimation method for estimating diffusion models. However, the likelihood estimation method for diffusion is complicated and needs much amount of computing. In this paper we develop the estimation methods which are simple enough to be compared to the Euler approximation method, and efficient enough statistically to be compared to the likelihood estimation method. We devise pseudo-likelihood and propose the maximum pseudo-likelihood estimation methods. The pseudo-likelihoods are obtained by approximating the transition density with normal distributions. The means and the variances of the distributions are obtained from the delta expansion suggested by Lee, Song and Lee (2012). We compare the newly suggested estimators with other existing estimators by simulation study. From the simulation study we find the maximum pseudo-likelihood estimator has very similar properties with the maximum likelihood estimator. Also the maximum pseudo-likelihood estimator is easy to apply to general diffusion models, and can be obtained by simple numerical steps.

• Communications for Statistical Applications and Methods
• /
• v.14 no.1
• /
• pp.229-239
• /
• 2007

We present a diagnostic method for the quasi-likelihood estimators in generalized linear models. Since these estimators can be usually obtained by iteratively reweighted least squares which are well known to be very sensitive to unusual data, a diagnostic step is indispensable to analysis of data. We extend the local influence approach based on the maximum likelihood function to that on the quasi-likelihood function. Under several perturbation schemes local influence diagnostics are derived. An illustrative example is given and we compare the results provided by local influence and deletion.

• Communications for Statistical Applications and Methods
• /
• v.14 no.2
• /
• pp.309-316
• /
• 2007

In this paper, we decompose the whole likelihood based on grouped data into conditional likelihoods and study the approximate contribution of additional inspection to the efficiency. We also combine the conditional maximum likelihood estimators to construct an approximate maximum likelihood estimator. For an exponential distribution, we see that a large inspection size does not increase the efficiency much if the failure rate is small, and the maximum likelihood estimator can be approximated with a linear function of inspection times.