Quantcast

Testing for changes in polynomial regression

Research paper by Alexander Aue, Lajos Horváth, Marie Hušková, Piotr Kokoszka

Indexed on: 22 Oct '08Published on: 22 Oct '08Published in: Mathematics - Statistics



Abstract

We consider a nonlinear polynomial regression model in which we wish to test the null hypothesis of structural stability in the regression parameters against the alternative of a break at an unknown time. We derive the extreme value distribution of a maximum-type test statistic which is asymptotically equivalent to the maximally selected likelihood ratio. The resulting test is easy to apply and has good size and power, even in small samples.