Nonparametric tests in linear model with autoregressive errors

Abstrakt

In the linear regression model with possibly autoregressive errors, we construct a family of nonparametric tests for significance of regression, under a nuisance autoregression of model errors. The tests avoid an estimation of nuisance parameters, in contrast to the tests proposed in the literature.

A simulation study illustrate their good performance.

Klíčová slova

• Autoregression
• Autoregression rank scores
• Linear regression
• Rank test
• Regression rank scores