similar to: cv.glm: delta squared --> squared q

Hi! When I use the glm.cv function I get a value called "delta" which is explained as the "raw cross-validation estimate of prediction error". I recently found a formula for that term in literature where it is defined as: alpha = 1 / N * sum over( yi - yi,pred,CV) Well it is somehow similar to the RSS for R2 and the PRESS for Q2. But this delta value increases with

Hi, everyone, I have a question regarding function cv.glm in library 'boot'. Basically cv.glm can calculate the estimated K-fold cross-validation prediction error for generalized linear models. My question is this: if I am fitting a logit model, what kind of threshold will it use to calculate the prediction error (saved in 'delta')? It will use 0.5 as the threshold or pick a

I have a couple quick questions about the use of cv.glm for cross-validation. 1. If we have a Poisson GLM with counts Y~Poisson(mu) and ln(mu)=beta0+beta1*x1+..., is the prediction error (delta) that is output from cv.glm provided in terms of the counts (y) or the (mu)? 2. Can cv.glm be used for negative binomial models fit using glm.nb? It appears to work, but since NB models aren't

Hello, I used the cv.lm function to validate a linear regression model fit<-lm(y ~ x1+x2+x3+x4+0, data=mydata) without intercept I tried to validate the model by performing a leave one out cross validation procedure usinfg the cvlm function: CVlm(df=mydata, fit, m=196) But how can I get the adjusted R? from the output of this function. Or is there any other function to

I am trying to write a wrapper for the last example in help(cv.glm) that deals with leave-one-out-cross-validation (LOOCV) for a logistic model. This wrapper will be used as part of a bigger program. Here is my wrapper funtion : logistic.LOOCV.err <- function( formu=NULL, data=NULL ){ cost.fn <- function(cl, pred) mean( abs(cl-pred) > 0.5 ) glmfit <- glm(

Hi, everyone When I was using cv.lm(DAAG) , I found there might be something wrong with it. The problem is that we can't use it to deal with a linear model with more than one predictor variable. But the usage documentation hasn't informed us about this. You can find it by excuting the following code: xx=matrix(rnorm(20*3),ncol=3) bb=c(1,2,0) yy=xx%*%bb+rnorm(20,0,10)

Hi, I've done some searching where others have had trouble with this error (or "warning" actually), but I'm unable to solve my problem. I have a data sheet with 13 columns and 36 rows. Each column has exactly the same number of rows. I've created glms and now want to do cross-validation on 2 of them. Please be gentle-- I'm new to R (and statistics, too, for that

Dear list members. I have problems with the usage of cv.glm from the boot package. Here are some parts of the script I wanted to use: data <- read.table("selected_2D.csv", header=TRUE, sep=",") ? glm.fitted <- glm("ydata$ y ~ 1 + density + vsurf_ID6 + vsurf_S ", data=data) error <- cv.glm(data=data, glm.fitted, K=6) ydata$y is a separate data set, where

I have a logistic model fitted with the following R function: glmfit<-glm(formula, data, family=binomial) A reasonable cutoff value in order to get a good data classification (or confusion matrix) with the fitted model is 0.2 instead of the mostly used 0.5. And I want to use the `cv.glm` function with the fitted model: cv.glm(data, glmfit, cost, K) Since the response in the fitted

Dear list, I am modelling fish distributions using the glm-function followed by the step-function, and then want to cross-validate the model via the cv.glm-function from the {boot} package. I am working on fish distributions on coral reefs. The code I have works for one fish species. Since I have 227 fishes, I wrote a loop. Now the cv.glm-function comes up with an error message: "Error in

Hi, I have a question regarding the cv.glm function in the package boot. What is exactly the cost? Is it the threshold value for an estimated value to be classified as either 0 or 1? I have troubles understanding the explanation in R. Lets say I want all estimations >0.65 to be classified as 1s and <0.35 as 0s, how do I write that? And if the cost is something else, how do I set the

Hey all, I need some help with a cross validation. I'm new with R and as well with statistics. I had a group work to create a tool for remote sensing class that extracts the best bands of hyperspectral satellite images that describe vegetation. Its a regression between a linear function of using a normalized differenced index (i-j)/(i+j) while i and j are the bands (in the data these are the

Georges Monette and I would like to announce a new package, cv, now on CRAN, which implements cross-validation of regression models. Some of the functions supplied by the package: - cv() is a generic function with a default method and computationally efficient "lm" and "glm" methods, along with a method for a list of competing models. There are also experimental

Georges Monette and I would like to announce a new package, cv, now on CRAN, which implements cross-validation of regression models. Some of the functions supplied by the package: - cv() is a generic function with a default method and computationally efficient "lm" and "glm" methods, along with a method for a list of competing models. There are also experimental

I am try to cross validate some logistic regressions. cv.glm allows me to do this randomly but since I have data over a number of years and over a number of distince areas, I would like to cross-validate temporarly and spatially. I've already attached the temporal and spatial attributes to my data, but I'm unsure as to how to achieve this, as it seems I can't use cv.glm for this

Dear R-list, I'm not sure what I've found about a function in DAAG package is a bug. When I was using cv.lm(DAAG) , I found there might be something wrong with it. The problem is that we can't use it to deal with a linear model with more than one predictor variable. But the usage documentation hasn't informed us about this. The code illustrates my discovery: > library(DAAG)

Hi. Is there a way to get the values predicted from (leave-one-out) cv.glm? It seems like a useful diagnostic to plot observed vs. predicted values. Thanks, Jeff **************************************** Jeffrey A. Stratford, Ph.D. Postdoctoral Associate 331 Funchess Hall Department of Biological Sciences Auburn University Auburn, AL 36849 334-329-9198 FAX 334-844-9234

Dear all, I am making model selection of generalised linear models based on delta AIC, using the command step(AIC). Also, I would like to learn the explanatory powers of each independent variable. To phrase differently, it is in need to show the delta AIC of all models rather than the final model with the least delta AIC. Please kindly advise if it is possible to exhibit all delta AIC of

Hi All, I have a fitted model called glm.fit which I used glm and data dat is my data frame pred= predict(glm.fit, data = dat, type="response") to predict how it predicts on my whole data but obviously I have to do cross-validation to train the model on one part of my data and predict on the other part. So, I searched for it and I found a function cv.glm which is in package boot.

Dear R-Helpers, I am trying to bootstrap the coefficient of variation on a suite of vectors, here I provide an example using one of the vectors in my study. When I ran this script with the vector x <-c(0.625, 0.071428571, 0.133333333, 0.125, 0), it returned CV(boot) [the second one], and stderr(boot) [the second one] without problem. However, when I ran it with the vector in the