Sorry that I forgot to include R Help in the addressee list. Here is one of my earlier follow-up emails.
---------- Forwarded message ---------- From: Xiaogang Su <xiaogan...@gmail.com> Date: Wed, Feb 26, 2014 at 8:48 AM Subject: Re: [R] Fitting glm with maxit=0? To: Prof Brian Ripley <rip...@stats.ox.ac.uk> Thanks Prof. Ripley for the reply. Certainly, the theorem is for MLE. But MLE is not the only best asyptotical normal estimator. In particular, i am interested in a class of estimators that are close to MLE in an order smaller than 1/sqrt(n). By Slutsky theorem, they also follow the same asymptotic normal dist as MLE. In computing the information matrix, one could use the estimate itself instead of plugging in MLE. Anyhow, i think it would be a good feature to have with glm. One last note, forcing at least one more iteration, as in the current version of glm, is not wrong as that essentially gives a one step estimator. ============================= Xiaogang Su, Ph.D. Associate Professor Department of Mathematical Sciences University of Texas at El Paso 500 W. University Ave. El Paso, Texas 79968-0514 x...@utep.edu xiaogan...@gmail.com https://sites.google.com/site/xgsu00/ On Feb 26, 2014 1:36 AM, "Prof Brian Ripley" <rip...@stats.ox.ac.uk> wrote: > The theory used assumes that the estimates are MLEs (of the linear > predictor). > > One could say that > > the variance-covariance matrix for any arbitrarily given estimates > > is zero: there is no variability. > > On 26/02/2014 08:24, Xiaogang Su wrote: > >> Dear All, >> >> Does anyone know if there is any way to obtain the variance-covariance >> matrix for any arbitrarily given estimates with the glm() function? >> >> Here is what I really want. Given an arbitrary estimate (i.e., as starting >> points with the start= argument), the glm() function could return the >> corresponding variance-covariance matrix (or Hessian) and other quantities >> with no Netwon-Raphson iteration? This could have been done by setting >> maxit=0, but very unfortunately that is not an option in glm.control(). To >> illustrate the problem, >> >> mydata <- read.csv("http://www.ats.ucla.edu/stat/data/binary.csv") >> beta0 <- 1:3 >> control0 <- glm.control(epsilon = 1e10, maxit = 0, trace = FALSE) >> fit <- glm(admit ~ gre + gpa, data = mydata, family = "binomial", >> start=beta0, control=control0) >> summary(fit)$"cov.scaled" >> >> By the way, I am aware that I could directly compute the covariance matrix >> using the formula. I also know that I could extract the corresponding >> deviance by using the offset option. >> >> Any suggestion is greatly appreicated. >> >> Thanks, >> Xiaogang Su >> >> ============================= >> Xiaogang Su, Ph.D. >> Associate Professor >> Department of Mathematical Sciences >> University of Texas at El Paso >> 500 W. University Ave. >> El Paso, Texas 79968-0514 >> x...@utep.edu >> xiaogan...@gmail.com >> https://sites.google.com/site/xgsu00/ >> >> [[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. >> >> > > -- > Brian D. Ripley, rip...@stats.ox.ac.uk > 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 > [[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.