specifies whether result should be centered (default), left-aligned or right-aligned.
endrule
character string indicating how the values at the
beginning and the end (of the data) should be treated.
"keep"
keeps the first and last \(k_2\) values
at both ends, where \(k_2\) is the half-bandwidth k2
= k %/% 2,
i.e., y[j] = x[j] for \(j \in \{1,\ldots,k_2;
n-k_2+1,\ldots,n\}\);
"constant"
fill the ends with first and last calculated value in output array
(out[1:k2] = out[k2+1])
"NA"
the default, leaves the values to NA, as they are returned by filter.
Value
Returns a vector of the same size and same class as x.
Details
The implementation is using the function filter to calculate the moving average.
See Also
There's a faster implementation of running mean in the package caToolsrunmean() and a slower one in forecastma(). There's similar code in Midx().