A Permutation Approach to Goodness-of-fit Testing in Linear Regression Models (original) (raw)

2019

Model checking plays an important role in linear regression as model misspecification seriously affects the validity and efficiency of regression analysis. In practice, model checking is often performed by subjectively evaluating the plot of the model’s residuals. This approach is objectified by constructing a random process from the model’s residuals, however due to a very complex covariance function obtaining the exact distribution of the test statistic is intractable. Several solutions to overcome this have been proposed, however the simulation and bootstrap based approaches are only asymptotically valid and can, with a limited sample size, yield tests which have inappropriate size. We therefore propose to estimate the null distribution by using permutations. We show, under some mild assumptions, that with homoscedastic random errors this yields consistent tests under the null and the alternative hypotheses. Small sample properties of the proposed tests are studied in an extensiv...

Permutation testing in high-dimensional linear models: an empirical investigation

Journal of Statistical Computation and Simulation, 2020

Permutation testing in linear models, where the number of nuisance coefficients is smaller than the sample size, is a well-studied topic. The common approach of such tests is to permute residuals after regressing on the nuisance covariates. Permutation-based tests are valuable in particular because they can be highly robust to violations of the standard linear model, such as non-normality and heteroscedasticity. Moreover, in some cases they can be combined with existing, powerful permutation-based multiple testing methods. Here, we propose permutation tests for models where the number of nuisance coefficients exceeds the sample size. The performance of the novel tests is investigated with simulations. In a wide range of simulation scenarios our proposed permutation methods provided appropriate type I error rate control, unlike some competing tests, while having good power.

Permutation Test for a Multiple Linear Regression Model

2006

To test the partial regression coefficients in a multiple linear regression model, a nonparametric permutation test based on the partial F statistic is proposed and, when sample size is large, it is significantly as good as the partial F test even in the case of normal distribution. This test can be applied to any distribution of errors. Simulation results show that, for non-normal distributions and in term of power, the proposed test performs quite well compared with the partial F test and the other permutation tests.

A Monte Carlo permutation procedure for testing variance components in generalized linear regression models

Computational Statistics

Testing zero variance components is of utmost importance in various applications empowered by the use of mixed-effects models. Focusing on generalized linear models, this article proposes a permutation test using an analogue of the ANOVA test statistic that merely requires fitting the null model with independent observations. Monte Carlo simulations reveal that the new test has correct Type-I error rate and that its power compares favorably to an existing bootstrap score test. A real data application illustrates the advantageous capability of the proposed test in detecting the need for random effects.

Loading...

Loading Preview

Sorry, preview is currently unavailable. You can download the paper by clicking the button above.