[R] SVAR: error message

William Dunlap wdun|@p @end|ng |rom t|bco@com
Tue Oct 23 22:14:30 CEST 2018


The immediate problem is in svartype=="AB-model" part of SVAR:

            pos <- which(is.na(rb))
            cols <- length(pos)
            Rb <- matrix(0, nrow = Ksq, ncol = cols)
            for (i in 1:cols) Rb[pos[i], i] <- 1

Your Bmat has no NA's, so cols is 0.  Then 1:cols is c(1,0), causing
an error in the for loop when it tries to reference column 1 of a 0-column
matrix.  The code should use seq_len(cols) instead of 1:cols (or give
an error telling the user that Bmat should contain some NA's).

This is the sort of question one should send to the maintainer of the
package
that SVAR is in:
>  maintainer("vars")
[1] "Bernhard Pfaff <bernhard using pfaffikus.de>"



Bill Dunlap
TIBCO Software
wdunlap tibco.com

On Mon, Oct 22, 2018 at 8:15 PM, John <miaojpm using gmail.com> wrote:

> Hi,
>
>   I have a bi-variate VAR model and would like to convert it to SVAR but
> get an error message. Could someone pinpoint anything wrong and correct my
> code? Thanks,
>
> John
>
>
> amat <- matrix(c(NA, 0, NA, NA), nrow = 2, ncol = 2, byrow = TRUE)
> df1<-data.frame(x=c(1,4,5,6,7,8,9,3,5,3), y=c(4,7,1,2,3,9,3,4,5,7))
> var1<-VAR(df1, type = "const", ic="BIC")
> svar2<-SVAR(x = var1, estmethod = "scoring", Amat = amat, Bmat = diag(2),
>             max.iter = 100, conv.crit = 0.1e-6, maxls = 1000)
>
> > amat <- matrix(c(NA, 0, NA, NA), nrow = 2, ncol = 2, byrow = TRUE)
> > df1<-data.frame(x=c(1,4,5,6,7,8,9,3,5,3), y=c(4,7,1,2,3,9,3,4,5,7))
> > var1<-VAR(df1, type = "const", ic="BIC")
> > svar2<-SVAR(x = var1, estmethod = "scoring", Amat = amat, Bmat = diag(2),
> +             max.iter = 100, conv.crit = 0.1e-6, maxls = 1000)
> Error in `[<-`(`*tmp*`, pos[i], i, value = 1) : subscript out of bounds
>
>         [[alternative HTML version deleted]]
>
> ______________________________________________
> R-help using r-project.org mailing list -- To UNSUBSCRIBE and more, see
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/
> posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>

	[[alternative HTML version deleted]]




More information about the R-help mailing list