similar to: variance specification using glm and quasi

It seems that MASS suggest to judge on the basis of sum(residuals(mode,type="pearson"))/df.residual(mode). My question: Is there any rule of thumb of the cutpoiont value? The paper "On the Use of Corrections for Overdispersion" suggests overdispersion exists if the deviance is at least twice the number of degrees of freedom. Are there any further hints? Thanks. -- Ronggui

Dear All, I am using the function vuong from pscl package to compare 2 non nested models NB1 (negative binomial I ) and Zero-inflated model. NB1 <-  glm(, , family = quasipoisson), it is an object of class: "glm" "lm" zinb <- zeroinfl( dist = "negbin") is an object of class: "zeroinfl"   when applying vuong function I get the following: vuong(NB1,

Hi R, This is a very trivial one.... C=0.1 I want to check whether my value of C is between 0 and 1 exclusively.... I don't want to use (C>0 & C<1). And I can't use a single statement like (0<C<1). Is there a between function? Or how do we specify from 0 to 1? Does %in% help me? Many Thanks, Shubha This e-mail may contain confidential and/or

Can anyone explain to me why the AIC values are so different when using glm.nb and glm with a negative.binomial family, from the MASS library? I'm using R 1.8.1 with Mac 0S 10.3.4. >library(MASS) > dfr <- data.frame(c=rnbinom(100,size=2,mu=rep(c(10,20,100,1000),rep(25,4))), + f=factor(rep(seq(1,4),rep(25,4)))) > AIC(nb1 <- glm.nb(c~f, data=dfr)) [1] 1047 >

Hello everyone, Sorry if my question is not clear, my first language is not English, but Portuguese. I am building a model for my data, using non-binomial error. I am having a bit of a problem when updating the model to remove parameters that I no do no autocorrelate with other variables (I have used a autocorrelation function for this). So my first model looks like this:

Hi all, This is a very silly question or something escapes me: Let obj a simple gam poisson model. Let >obj<-gam(....,family=poisson) >obj1<-update(obj, family=quasi(link="log", var="mu")) >From summary.glm(obj1) the dispersion parameter is estimated 1.165; In fact it is: > (predict(obj1, se.fit=T)$se.fit[1:5]/predict(obj, se.fit=T)$se.fit[1:5])^2 4

I get the following error when I run qplot() qplot(grade, read,data = hhm.long.m, geom = c("point", "smooth")) Error in smooth.construct.cr.smooth.spec(object, data, knots) : x has insufficient unique values to support 10 knots: reduce k. I am not sure how to tackle this problem. When I take a subsample (< 1000) than I am able to run that function but with my sample

How might you fit a generalized linear model (glm) with variance = mu+theta*mu^2 (where mu = mean of the exponential family random variable and theta is a parameter to be estimated)? This appears in Table 2.7 of Fahrmeir and Tutz (2001) Multivariate Statisticial Modeling Based on Generalized Linear Models, 2nd ed. (Springer, p. 60), where they compare "log-linear model fits to

Hhm. OK, but it doesn't work. The ups goes offline and starts back immediately. Even if ups.delay.start is set 0. I will try another test circuit after using the upsrw command. -----Ursprüngliche Nachricht----- Von: Nut-upsuser Im Auftrag von Roger Price Gesendet: Sonntag, 27. Oktober 2019 11:11 An: Nut-upsuser at alioth-lists.debian.net Betreff: Re: [Nut-upsuser] Problems with NUT on

[This email is either empty or too large to be displayed at this time]

Dear R community, I would like to compare the degree of aggregation (or dispersion) of bacteria isolated from plant material. My data are discrete counts from leaf washes. While I do have xy coordinates for each plant, it is aggregation in the sense of the concentration of bacteria in high density patches that I am interested in. My attempt to analyze this was to fit negative binomial

Full_Name: George Russell Version: 2.10.0 OS: Microsoft Windows XP Service Pack 2 Submission from: (NULL) (217.111.3.131) The following code (typed into R --vanilla) testString <- '"B1\nB2"\n1\n' con <- textConnection(testString) tab <- read.csv(con,stringsAsFactors = FALSE) produces a data frame with with one row and one column; the name of the column is

Hello, I have a question about the add1() function and quasilikelihoods for GLMs. I am fitting quasi-Poisson models using glm(, family = quasipoisson). Technically, with the quasilikelihood approach the deviance does not have the interpretation as a likelihood-based measure of sample information. Functions such as stepAIC() cannot be used. The function add1() returns the change in the scaled

Hello, I have the following r-codes for solving a quasilikelihood estimating equation: >library(geepack) >fit<-geese(y~x1+x2+x3,jack=TRUE,id=id,scale.fix=TRUE,data=dat,mean.link = "logit", corstr="independence") Now my question is how can I calculate the confidence interval of the parameters of the above model "fit"? [[alternative HTML version deleted]]

Hi, I recently found out that I was able to get system mails, but not mails from the outside world. What setting do I have to fiddle with for this? Regards,

Full_Name: Robert Kushler Version: 2.7.2 OS: Windows XP Submission from: (NULL) (69.246.102.98) I believe that the negative binomial family (from MASS) should be added to the list for which dispersion is set to 1.

Hi Roger, it was set by ups.conf. Should I use the upsr-command? Bjoern -----Ursprüngliche Nachricht----- Von: Nut-upsuser Im Auftrag von Roger Price Gesendet: Sonntag, 27. Oktober 2019 11:06 An: nut-upsuser at alioth-lists.debian.net Betreff: Re: [Nut-upsuser] Problems with NUT on Raspberry Pi On Sun, 27 Oct 2019, EP wrote: > OK, reporting back. Just tested the setting

Same behavior here. I tested it by executing the usrw-command (bash) as well. UPS turns off and on again. It even seems to ignore the ups.delay.shutdown (was set to 120 seconds but shutted down faster). Afterwards I set the ups.delay.start to -1. That makes the ups staying off. But - and that is the big mistake - it doesn't turn on when power returns with this setting. I can't imagine

Hello I am currently working on forecasting hourly electricity prices where I am required to do wavelet transformation. First I need to calculate decomposition coefficients of the wavelet transform, then I need to select wavelet function using multiresolution technique. I would be really grateful if I can find any help in this matter. Thanking You with regards, -- Name : Tarun Trivedi Senior

I have been asked to look at options for doing relative risk regression on some survey data. I have a binary DV and several predictor / adjustment variables. In R, would this be as "simple" as using the survey package to set up an appropriate design object and then running svyglm with family=binomial(log) ? Any other suggestions for covariate adjustment of relative risk