摘要:

In this paper, we analyze the rates of convergence of the mean square error of a partial spline estimator and a Denby/Speckman - type estimator for the parametric component of a semiparametric model under a wide range of conditions. It is found that the Denby/Speckman - type estimator has a faster rate of convergence than the partial spline estimator for some cases. Itis shown that the optimal rate of decrease for the smoothing parameter X under the criterion of minimizing MSE for the parametric component is not necessarily the same as the optimal rate for minimization of the predictive mean square error in the function estimate. Thus data based estimates for X optimal for predictive mean square error in the function, such as GCV, maynot be optimal for mean square error in the parametric component, leaving open the question of a data based estimate for X in this context.

ISSN:0361-0918    

DOI:10.1080/03610918808812716

出版商:Marcel Dekker, Inc.    

年代:1988

数据来源: Taylor

摘要:

The resubstitution estimator of classification error rates is known to have both an optimistic bias and a large variance. Modifications to this method have addressed these problems. the bootstrap estimator, for example, uses a resampling scheme to reduce bias, and the NS method uses a smoothing algorithm to reduce variance. In this paper we show that the use of a bootstrap adjustment to reduce the bias of the NS method results in an estimator which combines the advantages of small bias with low variance, and is therefore preferable to existing resampling estimators. In addition, a new smoothed estimator with reduced bias is introduced which may eliminate the need for resampling in somesituations.

ISSN:0361-0918    

DOI:10.1080/03610918808812717

出版商:Marcel Dekker, Inc.    

年代:1988

数据来源: Taylor

摘要:

The small sample performance of Zeger and Liang's extended generalized linear models for the analysis of longitudinal data (Biometrics, 42,121-130,1986) is investigated for correlated gamma data. Results show that the confidence intervals do not provide desirable coverage of the true parameter due to considerably biased point estimates. Improved estimates are proposed using the jackknife procedure. Simulations performed to evaluate the proposed estimates indicate superior properties to the previous estimates.

ISSN:0361-0918    

DOI:10.1080/03610918808812718

出版商:Marcel Dekker, Inc.    

年代:1988

数据来源: Taylor

摘要:

Ten different estimators of the parameter in a limiting or serial dilution assay are compared. Eight of them are constructed to reduce the bias of the commonly used maximum likelihood estimator. Extensive Monte Carlo experiments using various designs, and practical considerations, suggest that a particular jackknife version of the maximum likelihood estimator is preferred, provided that the design is not too small.

ISSN:0361-0918    

DOI:10.1080/03610918808812719

出版商:Marcel Dekker, Inc.    

年代:1988

数据来源: Taylor

摘要:

A recursive formula for evaluating incomplete beta integral is obtained by reducing both parameters simultaneously. Applying this identity, we derive an exact expression for F-cumulative probabilities. An algorithm is then derived by using this expression. Results from a single-precision FORTRAN program by using this algorithm are compared with that using MDFD subroutine of IMSL in both of accuracy and speed for degrees of freedom 1 lsequal; m, n $lsequal; 500. It is seen that both results coincide with at least 11 significant digits, and the new algorithm is consistently faster than MDFD by 10 to 15 percent.

ISSN:0361-0918    

DOI:10.1080/03610918808812720

出版商:Marcel Dekker, Inc.    

年代:1988

数据来源: Taylor

摘要:

The Kolmogorov-Smirnov (KS) test is an empirical distribution function (EDF) based goodness-of-fit test that requires the underlying hypothesized density to be continuous and completely specified. When the parameters are unknown and must be estimated from the data, standard tables of the KS test statistic are not valid. Approximate upper tail percentage points of the KS statistic for the inverse Gaussian (IG) distribution with unknown parameters are tabled in this paper.

ISSN:0361-0918    

DOI:10.1080/03610918808812721

出版商:Marcel Dekker, Inc.    

年代:1988

数据来源: Taylor

摘要:

The purpose of this paper is to compare sampling performance of the minimax generalized ridge regression estimators considered by Casella (1985) with that of ordinary least squares estimator by numerical calculations of exact mean squared error of these estimators.

ISSN:0361-0918    

DOI:10.1080/03610918808812722

出版商:Marcel Dekker, Inc.    

年代:1988

数据来源: Taylor

摘要:

The variance of the Maximum Likelihood Estimator (MLE) of the slope parameter in a logistic regression model becomes large as the degree of collinearity among the explanatory variables increases. In a Monte Carlo study, we observed that a ridge type estimator is at least as good as, and often much better than, the MLE in terms of Total and Prediction Mean Squared Error criteria. Using a set of medical data it is illustrated that the ridge trace of the estimator considered here is a useful diagnostic tool in logistic regression analysis.

ISSN:0361-0918    

DOI:10.1080/03610918808812723

出版商:Marcel Dekker, Inc.    

年代:1988

数据来源: Taylor

摘要:

The theory and properties of trend-free (TF) and nearly trend-free (NTF) block designs are wel1 developed. Applications have been hampered because a methodology for design construction has not been available.

ISSN:0361-0918    

DOI:10.1080/03610918808812724

出版商:Marcel Dekker, Inc.    

年代:1988

数据来源: Taylor

摘要:

Monte Carlo methods are used to examine the small-sample properties of 11 test statistics that can be used for comparing several treatments with respect to their mortality experiences while adjusting for covariables. The test statistics are investigated from three distinct models: the parametric, semiparametric and rank analysis of covariance (Quade, 1967) models. Four tests (likelihood ratio, Wald, conditional and unconditional score tests) from each of the first two models and three tests (based on rank scores) from the last model are discussed. The empirical size and power of the tests are investigated under a proportional hazards model in three situations: (1) the baseline hazard is correctly assumed to be Exponential, (2) the baseline hazard is incorrectly assumed to be Exponential, and (3) a treatment-covariate interaction is omitted from the analysis.

ISSN:0361-0918    

DOI:10.1080/03610918808812725

出版商:Marcel Dekker, Inc.    

年代:1988

数据来源: Taylor