anova.bvpot {POT} | R Documentation |
Computes analysis of deviance for ``bvpot'' object
## S3 method for class 'bvpot': anova(object, object2, ..., half = FALSE)
object, object2 |
Two objects of class ``bvpot'', most often return of the
fitbvgpd function. |
... |
Other options to be passed to the anova
function. |
half |
Logical. For some non-regular testing problems the deviance
difference is known to be one half of a chi-squared random
variable. Set half to TRUE in these cases. |
This function returns an object of class anova. These objects represent analysis-of-deviance tables.
Circumstances may arise such that the asymptotic distribution of the test statistic is not chi-squared. In particular, this occurs when the smaller model is constrained at the edge of the parameter space. It is up to the user recognize this, and to interpret the output correctly.
In some cases the asymptotic distribution is known to be one half of a
chi-squared; you can set half = TRUE
in these cases.
Mathieu Ribatet (Alec Stephenson for the ``Warning'' case)
x <- rgpd(1000, 0, 1, -0.25) y <- rgpd(1000, 2, 0.5, 0) M0 <- fitbvgpd(cbind(x,y), c(0, 2)) M1 <- fitbvgpd(cbind(x,y), c(0,2), model = "alog") anova(M0, M1) ##Non regular case M0 <- fitbvgpd(cbind(x,y), c(0, 2)) M1 <- fitbvgpd(cbind(x,y), c(0, 2), alpha = 1) anova(M0, M1, half = TRUE)