On Fri, 21 Apr 2023 09:02:37 +0200 Patrick Giraudoux <patrick.giraud...@univ-fcomte.fr> wrote:
> I meet an error with glm.nb that I cannot explain the origin (and > find a fix). The model I want to fit is the following: > > library(MASS) > > glm.nb(deg~offset(log(durobs))+zone,data=db) > > and the data.frame is dumped below. Thank you for providing both the code and a small piece of data that reproduces the error! (It almost worked. Your mailer automatically generated a plain text version of the e-mail and put Unicode non-breaking spaces in there. R considers it a syntax error to encounter any of the various Unicode space-like characters outside string literals.) > deg = structure(c(0, 1, 0, 3, 0, 1, 0, 2, 1, 0, 3, 0, 0, 0, 4, 1, 0, > 0, 0, 0, 4, 0, 0, 0, 4, 3, 2, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, > 0, 3, 2, 3, 0, 1, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 1, 0, 1, 2, > 0, 0, 0, 2, 1, 0, 2, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 2, > 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 2, 1, 0, 0, 1, 3, 2, > 1, 2, 0, 0, 0, 0, 0, 1, 0, 0, 2, 0, 0, 0, 0, 0, 1, 2, 0, 0, 0, 0, 2, > 1, 1, 0), dim = 135L, class = "table") The problem is that `deg` is a table, which ends up making the effective weights a table too. tables are arrays, and element-wise product rules are stricter for them than for plain matrices. The code makes use of the ability to take an element-wise product between a matrix and a vector of the same length as the number of rows in the matrix: matrix(1:12, 4) * 1:4 # works matrix(1:12, 4) * as.array(1:4) # results in the same error # the right way to take products with an array is to make sure that the # shapes match exactly matrix(1:12, 4) * as.array(cbind(1:4, 1:4, 1:4)) One possible solution is to to remove all attributes from db$deg: db$deg <- as.vector(db$deg) This way the values of the expressions involved in glm.fit end up being of the expected type, and the function completes successfully. -- Best regards, Ivan ______________________________________________ R-help@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.
