Hi, Milton, Thank you for the reply. I tried, but it seems the problem is the column name of the test data is not the same as the column name of the training data. I didn't give the column name, the system seemed do. How to chang here?
Annie On Fri, Aug 7, 2009 at 7:52 PM, milton ruser <milton.ru...@gmail.com> wrote: > Hi Annie, > > create a new data.frame with input variables having all predictors > variables on it. > after give a look at ?predict > > best wishes > > milton > > On Fri, Aug 7, 2009 at 8:19 PM, annie Zhang > <annie.zhang2...@gmail.com>wrote: > >> Hi, R users, >> >> I am trying to use glm to do logistic regression. I know generally when I >> have two covariates, say x1 and x2, then I do >> fit <- glm(y~x1+x2,famliy='binomial') >> But now my covariates form a n*p matrix, say x, so actually each column is >> a >> covariate. So I think I should do >> fit <- glm(y~x,family='binomial') >> Then I need to predict new data. How should I write the newdata? I tried >> several thing, all failed. The x in the fit is a matrix, but is a vector >> for >> the new data. >> >> Thank you, >> Annie >> >> [[alternative HTML version deleted]] >> >> ______________________________________________ >> R-help@r-project.org mailing list >> https://stat.ethz.ch/mailman/listinfo/r-help >> PLEASE do read the posting guide >> http://www.R-project.org/posting-guide.html<http://www.r-project.org/posting-guide.html> >> and provide commented, minimal, self-contained, reproducible code. >> > > [[alternative HTML version deleted]] ______________________________________________ R-help@r-project.org mailing list 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.
