In this paper a robust version of the classical Wald test statistics for linear hypothesis in logistic regression model is introduced and its properties are explored. We study problem under assumption random covariates although some ideas with non also considered. A family type tests considered here, where minimum density power divergence estimator used instead maximum likelihood estimator. obtain asymptotic distribution robustness these statistics. The investigated theoretically through...

In testing of hypothesis the robustness tests is an important concern. Generally, maximum likelihood based are most efficient under standard regularity conditions, but they highly non-robust even small deviations from assumed conditions. this paper we have proposed generalized Wald-type on minimum density power divergence estimators for parametric hypotheses. This method avoids use nonparametric estimation and bandwidth selection. The trade-off between efficiency controlled by a tuning...

This paper develops a new family of estimators, the minimum density power divergence estimators (MDPDEs), for parameters one-shot device model as well test statistics, Z-type statistics based on MDPDEs, testing corresponding parameters. The MDPDEs contains particular case maximum likelihood estimator (MLE) considered in Balakrishnan and Ling (2012). Through simulation study, it is shown that some have better behavior than MLE terms robustness. At same time, can be seen tests classical Z-test...

In regression models involving economic variables such as income, log transformation is typically taken to achieve approximate normality and stabilize the variance. However, often interest predicting individual values or means of variable in original scale. Under a nested error model for target variable, we show that usual approach back transforming predicted may introduce substantial bias. We obtain optimal (or "best") predictors small area under model. Empirical best are defined by...

Classical inferential methods for one-shot device testing data from an accelerated life-test are based on maximum likelihood estimators (MLEs) of model parameters. However, the lack robustness MLE is well-known. In this article, we develop robust by assuming a Weibull distribution as lifetime model. Wald-type tests these also developed. Their properties evaluated both theoretically and empirically, through extensive simulation study. Finally, inference proposed applied to three numerical...

Summary This article derives a new family of estimators, namely the minimum density power divergence as robust generalization maximum likelihood estimator for polytomous logistic regression model. Based on these Wald-type test statistics linear hypotheses is introduced. Robustness properties both proposed estimators and are theoretically studied through classical influence function analysis. Appropriate real life examples presented to justify requirement suitable statistical procedures in...

Abstract Introduced robust density‐based estimators in the context of one‐shot devices with exponential lifetimes under a single stress factor. However, it is usual to have several factors industrial experiments involving devices. In this paper, weighted minimum density power divergence (WMDPDEs) are developed as natural extension classical maximum likelihood (MLEs) for device testing data lifetime model multiple stresses. Based on these estimators, Wald‐type test statistics also developed....

AbstractThis paper presents a family of power divergence-type test statistics for testing the hypothesis elliptical symmetry. We assess performance new statistics, using Monte Carlo simulation. In this context, type I error rate as well tests are studied. Specifically, selected alternatives, we compare proposed procedure with that by Schott [Testing symmetry in covariance-matrix-based analyses, Stat. Probab. Lett. 60 (2002), pp. 395–404]. This last statistic is an easily computed one...

It sometimes occurs that one or more components of the data exert a disproportionate influence on model estimation. We need reliable tool for identifying such troublesome cases in order to decide either eliminate from sample, when collect was badly realized, otherwise take care use because results could be affected by components. Since measure detecting influential linear regression setting proposed Cook [Detection observations regression, Technometrics 19 (1977), pp. 15–18.], apart same...

This article presents new families of Rao-type test statistics based on the minimum density power divergence estimators which provide robust generalizations for testing simple and composite null hypotheses. The asymptotic distributions proposed tests are obtained their robustness properties also theoretically studied. Numerical illustrations provided to substantiate theory developed. On whole, seen be excellent alternatives classical Rao as well other well-known tests.

In this article, we develop robust estimators and tests for one-shot device testing under proportional hazards assumption based on divergence measures. Through a detailed Monte–Carlo simulation study numerical example, the developed inferential procedures are shown to be more against data contamination than classical procedures, maximum likelihood estimators.

In this paper, a robust version of the Wald test statistic for composite likelihood is considered by using minimum density power divergence estimator instead maximum estimator. This new family statistics will be called Wald-type statistics. The problem testing simple and null hypothesis considered, robustness studied on basis simulation study. also introduced, its asymptotic properties are studied.

In a recent article, Cardoso de Oliveira and Ferreira have proposed multivariate extension of the univariate chi-squared normality test, using known result for distribution quadratic forms in normal variables. this we propose family power divergence type test statistics testing hypothesis multinormality. The includes as particular case by Ferreira. We assess performance new Monte Carlo simulation. context, I error rates tests are studied, important members. Moreover, significant members...