similar to: how 'stepAIC' selects?

In a generalized linear model with k covariates, there are 2(kth power) - 1 possible models (excluding interactions). Awhile ago a posting to R-help suggested Model Selection and Multimodel Inference, 2nd ed, by Burnham and Anderson as a good source for understanding model selection. They recommend (page 71) computing AIC differences over all candidate models in the set of possible models. After

Dear Listers, I got this message when compiling a package: * creating pgirmess-manual.tex ... OK * checking pgirmess-manual.text ... ERROR LaTex errors when creating DVI version. This typically indicates Rd problems. The message is quite explicit but I struggled a lot before understanding that the trouble comes from a single file "selMod.rd" among 44 topics. Even though I have

Dear all, does anyone have r code to perform model-averaging of regression parameters by Akaike weights, and/or to do all-possible-subsets lm modelling that reports parameter estimates, AICc and number of parameters for each model? I have been looking for these in the archive but found none. (I am aware that many of you would warn me against these methods advocated by Burnham and Anderson

Do any R functions incorporate a sample sample size correction (e.g., Burnham and Anderson 1998). Thanks, Hank Stevens Martin Henry H. Stevens, Assistant Professor 338 Pearson Hall Botany Department Miami University Oxford, OH 45056 Office: (513) 529-4206 Lab: (513) 529-4262 FAX: (513) 529-4243 http://www.cas.muohio.edu/botany/bot/henry.html http://www.muohio.edu/ecology

oops, Forgot to cc to the list. Regards, Mike -----Original Message----- From: dr mike [mailto:dr.mike at ntlworld.com] Sent: 24 February 2005 19:21 To: 'Spencer Graves' Subject: RE: [R] Bayesian stepwise (was: Forward Stepwise regression based onpartial F test) Spencer, Obviously the problem is one of supersaturation. In view of that, are you aware of the following? A Two-Stage

This is a question at the border between stats and r. When I do a glm with many potential effects, and select a model using stepAIC, many independent variables are selected even if there are no relationship between dependent variable and the effects (all are random numbers). Do someone has a solution to prevent this effect ? Is it related to Bonferoni correction ? Is there is a ratio of

If AIC is giving you a model that is too large, then use BIC (log(n) as the penalty for adding a term in the model). This will yield a more parsimonious model. Now, if you ask me which is the better option, I have to refer you to the huge literature on model selection. Best, Ravi [[alternative HTML version deleted]]

More principled would be to use a lasso-type approach, which combines selection and estimation in one fell swoop! Ravi ________________________________ From: Ravi Varadhan Sent: Tuesday, June 6, 2017 10:16 AM To: r-help at r-project.org Subject: Subject: [R] glm and stepAIC selects too many effects If AIC is giving you a model that is too large, then use BIC (log(n) as the penalty for adding

Greetings, I have recently come into some confusion over weather or not AIC results for comparing among models requires that they be nested. Reading Burnham & Anderson (2002) they are explicit that nested models are not required, but other respected statisticians have suggested that nesting is a pre-requisite for comparison. Could anyone who feels strongly regarding either position

hi folks, I'm trying to build a table that contains information about a series of General Linear Models in order to calculate Akaike weights and other measures to compare all models in the series. i have an issue with indexing models and extracting the information (loglikehood, AIC's, etc.) that I need to compile them into the table. Below is some sample code that illustrates my

I am hoping to get some advise on the following: I am looking for an automatic variable selection procedure to reduce the number of potential predictor variables (~ 50) in a multiple regression model. I would be interested to use the forward stepwise regression using the partial F test. I have looked into possible R-functions but could not find this particular approach. There is a function

Dear list, I have been using multivariate multiple regression (MMR) in the form lm(Y~X) where Y and X are matrices of response and predictor variables. I know that summary(mlm.object) would give the usual lm statistics for each response variable separately and that anova.mlm(mlm.object) will give the analysis of variance table of the mlm object. However, anova.mlm (also manova(mlm.object))

A few general comments about stepwiseAIC and a suggestion of how to select models a) Apart form the problem, that stepwise selection is not a garanty to get the best model, you need to have a lot of data to avoid overfitting if your model includes 7 parameter plus interactions (> 10 observations per parameter is what you are ideally looking for). b) Have a look at Anderson and Burnham's

Dear listers, Here we have a strange result we can hardly cope with. We want to compare a null mixed model with a mixed model with one independent variable. > lmmedt1<-lme(mediane~1, random=~1|site, na.action=na.omit, data=bdd2) > lmmedt9<-lme(mediane~log(0.0001+transat), random=~1|site, na.action=na.omit, data=bdd2) Using the Akaike Criterion and selMod of the package pgirmess

Hello all, i hope you don't mind my off topic question. i want to use the Akaike criterion for variable selection in a regression model. Does anyone know some basic literature about that topic? Especially I'm interested in answers to the following questions: 1. Has (and if so how has) the criterion to be modified, if i estimate the transformations of the variables too? 2. How is the

Dear list, I gave up trying to fix my movie3d (rgl library) issue in my PC (completely black gif/png file) and went ahead and installed MacPorts and ImageMagick onto my iMac (OSX ver 10.6.8). I think ImageMagick is successfully installed in its default location (under /opt/local), and I ran movie3d but I get the following error: Error in system("convert --version", intern = TRUE) : ?

Dear list, I just recently installed R-2.13.0 on my Windows7. I used to run R-2.10.0. First of all, I used to be able to install R in English (in R-2.10.0) during the installation procedure. My PC is in a Japanese environment but I want R to be in English because I won't be able to interpret any errors if they are in Japanese (I am Japanese so I can read them but I won't be able to

R Users, I was just wondering if anyone has written a program (or if there is a package) out there that calculates the different derivations of AIC (e.g. AIC, AICc, QAIC, etc.) along with AIC differences (delta's), model likelihoods, Akaike weights and evidence ratio's (from Burnham and Anderson 2002). Just in a "for instance" if someone had the -2LL, sample sizes,

Dear all, I'm calculating matrix correlations with permutation tests and I got this funny result. All correlation coefficients are the same with mantel.test {ncf} and pcol {simba} but the two functions yield dramatically different p-values (using the same number of permutations). Could anyone please enlighten me what is causing the difference and which result I can trust? (My matrices

Dear list, I have a matrix Y of multiple response variables and a matrix X of predictor variables and I would like to fit a multivariate multiple regression model and compute the R2-value to determine the overall proportion of variance of the response matrix Y that is explained by the predictor matrix X. I have been using manova(Y ~ X) to assess the significance of the linear model. I am also