[R] The null hypothesis in kpss test (kpss.test())
Achim Zeileis
Achim.Zeileis at wu-wien.ac.at
Tue Mar 8 19:12:54 CET 2005
As help(kpss.test) tells you: kpss.test() approximates the p values by
interpolation from a simulated table of critical values. As p values
larger than 0.1 are typically regarded to be non-significant and p
values smaller than 0.01 are typically regarded to be highly
significant, the corresponding critical values are only stored for the
range 0.1 to 0.01.
Hence...
On Tue, 8 Mar 2005 12:51:37 -0500 (EST) Weiguang Shi wrote:
> is that 'x' is level or trend stationary. I did this
>
> > s<-rnorm(1000)
> > kpss.test(s)
>
> KPSS Test for Level Stationarity
>
> data: s
> KPSS Level = 0.0429, Truncation lag parameter = 7,
> p-value = 0.1
>
> Warning message:
> p-value greater than printed p-value in:
> kpss.test(s)
>
> My question is whether p=0.1 is a good number to
> reject N0?
...stationarity cannot be rejected here (which is not surprising) and...
> On the other hand, I have a series r and did the
> following:
> > plot.ts(r)
> > kpss.test(r)
>
> KPSS Test for Level Stationarity
>
> data: r
> KPSS Level = 3.1955, Truncation lag parameter = 7,
> p-value = 0.01
>
> Warning message:
> p-value smaller than printed p-value in:
> kpss.test(r)
...stationarity is clearly rejected here.
> So this says we can have more confidence in saying r
> is _not_ stationary?
Yes (I guess. I'm not sure about `more confidence'...`more' than what?)
Z
> Should I worry about the warnings?
>
> Thanks very much.
> Weiguang
>
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide!
> http://www.R-project.org/posting-guide.html
>
More information about the R-help
mailing list