[EMAIL PROTECTED] writes: > First, R-bugs is not for asking questions, only for reporting things you > are *certain* are bugs: see the R FAQ. > > Why are you using weights to omit cases? If you had used subset, this > would have worked. The problem is the use of zero weights, which are not > intended to be used in this way. We can fix this up, but it is not the > correct way to use glm.
I'm not even sure we want to fix this up. I recall some nasty issues with DF that have no proper solution that way - an observation with a tiny weight represents an observation with a large variance and contributes 1DF to the residual, with weight zero it is not supposed to contribute at all, so there's a discontinuity for weights approaching zero. > > On Tue, 28 Mar 2006, [EMAIL PROTECTED] wrote: > > > Full_Name: Robert Pusz > > Version: 2.2.1 > > OS: Windows > > Submission from: (NULL) (157.25.9.126) > > > > > > Hello, > > In my opinion something is wrong with 'weights' option in glm. > > My code is following: > > ###begin of the code#### > > cl <- c(5012, 106, 3410, 5655, 1092, 1513, 557, 1351, 3133, 2063, 3257, > > 4179, 5582, 5900, 8473, 4932, 3463, 5596, 2262, 0, 2638, 1111, > > 4881, 4211, 6271, 5257, 6926, 6165, 0, 0, 898, 5270, 2268, 5500, > > 6333, 1233, 1368, 0, 0, 0, 1734, 3116, 2594, 2159, 3786, 2917, > > 0, 0, 0, 0, 2642, 1817, 3479, 2658, 225, 0, 0, 0, 0, 0, 1828, > > 0, 649, 984, 0, 0, 0, 0, 0, 0, 599, 673, 603, 0, 0, 0, 0, 0, > > 0, 0, 54, 535, 0, 0, 0, 0, 0, 0, 0, 0, 172, 0, 0, 0, 0, 0, 0, > > 0, 0, 0) > > > > w <- c(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, > > 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 1, > > 1, 1, 1, 1, 1, 0, 0, 0, 0, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 1, 0, > > 1, 1, 0, 0, 0, 0, 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, > > 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0) > > > > row <- c(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, > > 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, > > 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, > > 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, > > 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10) > > > > col <- c(1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, > > 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 5, > > 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 7, 7, > > 7, 7, 7, 7, 7, 7, 7, 7, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 9, 9, 9, > > 9, 9, 9, 9, 9, 9, 9, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10) > > Row <- ordered(as.factor(row)) > > Col <- ordered(as.factor(col)) > > fit <- glm(cl ~ Row + Col, family = quasipoisson, weights = w) > > ###end of the code#### > > > > When I have written summary(fit) I have got: > > Estimate Std. Error t value Pr(>|t|) > > (Intercept) 7.64459 NA NA NA > > Row2 -0.05467 NA NA NA > > Row3 0.24541 NA NA NA > > Row4 0.42035 NA NA NA > > Row5 0.43961 NA NA NA > > Row6 0.04532 NA NA NA > > Row7 -0.04881 NA NA NA > > Row8 0.25370 NA NA NA > > Row9 -0.14976 NA NA NA > > Row10 -0.01267 NA NA NA > > Col2 0.69283 NA NA NA > > Col3 0.62603 NA NA NA > > Col4 0.27695 NA NA NA > > Col5 0.06056 NA NA NA > > Col6 -0.19582 NA NA NA > > Col7 -0.83044 NA NA NA > > Col8 -1.27914 NA NA NA > > Col9 -1.93235 NA NA NA > > Col10 -2.49709 NA NA NA > > > > When I omited 'weights=w' above table was filled in with numbers, but the > > results were wrong (because of taking zeros in regression). > > > > Could you tell me what's wrong? > > Kind regards, > > Robert > > > > ______________________________________________ > > R-devel@r-project.org mailing list > > https://stat.ethz.ch/mailman/listinfo/r-devel > > > > > > -- > Brian D. Ripley, [EMAIL PROTECTED] > Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/ > University of Oxford, Tel: +44 1865 272861 (self) > 1 South Parks Road, +44 1865 272866 (PA) > Oxford OX1 3TG, UK Fax: +44 1865 272595 > > ______________________________________________ > R-devel@r-project.org mailing list > https://stat.ethz.ch/mailman/listinfo/r-devel > -- O__ ---- Peter Dalgaard Øster Farimagsgade 5, Entr.B c/ /'_ --- Dept. of Biostatistics PO Box 2099, 1014 Cph. K (*) \(*) -- University of Copenhagen Denmark Ph: (+45) 35327918 ~~~~~~~~~~ - ([EMAIL PROTECTED]) FAX: (+45) 35327907 ______________________________________________ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel
