ADF Test using rollapply

Question

I am wondering if there is a way to do ADF Test on a given interval. Below is an example of me calculating the mean for every 20 samples:

rollapply(data, 20, FUN = mean)

I want to use the same logic and run ADF test for every 20 samples. Therefore I used the following code:

n = 24
test <- rollapply(spreadClose, n, FUN = adf.test)

Note: adf.test() is from library called "tseries".

Could anyone explain why this is not working? I am getting the following error.

Error in dimnames(x) <- dn : length of 'dimnames' [2] not equal to array extent

I would really appreciate any feedbacks!

Answer 1

I used the following codes and it started working:

adftest <- function(x)
{
    adfPvalue <-  adf.test(x)$p.value
    adfPvalue
}

PValue <- function(x) {
    Pvalue <- rollapply(x, N, adftest)
    Pvalue
}

PValue(out)

Answer 2

How about this:

library(zoo)
library(tseries)
z <- zoo(rnorm(100), as.Date(1:100))
rollapplyr(z, 20, adf.test)