Skip to content

Multiple time series as matrix

1 message · Giovanni Petris

Original
Giovanni Petris
# 
Hello,

I would like to do something like the following:

t(y) - FF %*% t(theta)

where y is a multiple time series, FF and theta are matrices. 
Unfortunately, this doesn't work. 

Using unclass(y) allows me to take the transpose, but when trying to
take the difference I get the
error "invalid time series parameters specified", I guess due to the
fact that t(unclass(y)) has still a tsp attribute. 

The following works:

t(matrix(y,NROW(y),NCOL(y))) - FF %*% t(theta)

But I was wondering if there is an easier way...

Thanks,
Giovanni