R help - Aug 2006 - what's wrong with my simulation programs on logistic regression

Dear friends,
 I'm doing a simulation on logistic regression model, but the programs
can't
work well,please help me to correct it and give some suggestions.
My programs:
data<-matrix(rnorm(400),ncol=8)  #sample size is 50
data<-data.frame(data)
names(data)<-c(paste("x",1:8,sep=""))  #8 independent
variables,x1-x8;
#logistic regression model is logit(y)=x1+x2+x3+x4+x5+x6+x7+x8
data$y<-exp(data$x1+data$x2+data$x3+data$x4+data$x5+data$x6+data$x7+data$x8)/(1+(data$x1+data$x2+data$x3+data$x4+data$x5+data$x6+data$x7+data$x8))

logist<-glm(y~.,family=binomial(),data=simdata)
*Warning messages:*
1: algorithm can't converge in: glm.fit(x = X, y = Y, weights = weights,
start = start, etastart = etastart,
2: the probability is 0 or 1 in: glm.fit (x = X, y = Y, weights = weights,
start = start, etastart = etastart,
-- 
With Kind Regards,
Zhi Jie,Zhang ,PHD
Department of Epidemiology
School of Public Health
Fudan University

	[[alternative HTML version deleted]]

Forgot to add my thinkings:
I think it over and think that the problem may be the
argument(data$y<-exp(data$x1+data$x2+data$x3+data$x4+data$x5+data$x6+data$x7+data$x8)/(1+(data$x1+data$x2+data$x3+data$x4+data$x5+data$x6+data$x7+data$x8))
),which maybe not correctly set for my
model:logit(y)=x1+x2+x3+x4+x5+x6+x7+x8,
Thanks very much!


On 8/31/06, zhijie zhang <epistat@gmail.com>
wrote:
>
>  Dear friends,
>  I'm doing a simulation on logistic regression model, but the programs
> can't work well,please help me to correct it and give some suggestions.
> My programs:
> data<-matrix(rnorm(400),ncol=8)  #sample size is 50
> data<-data.frame(data)
> names(data)<-c(paste("x",1:8,sep=""))  #8
independent variables,x1-x8;
> #logistic regression model is logit(y)=x1+x2+x3+x4+x5+x6+x7+x8
>
data$y<-exp(data$x1+data$x2+data$x3+data$x4+data$x5+data$x6+data$x7+data$x8)/(1+(data$x1+data$x2+data$x3+data$x4+data$x5+data$x6+data$x7+data$x8))
>
> logist<-glm(y~.,family=binomial(),data=simdata)
> *Warning messages:*
> 1: algorithm can't converge in: glm.fit(x = X, y = Y, weights =
weights,
> start = start, etastart = etastart,
> 2: the probability is 0 or 1 in: glm.fit (x = X, y = Y, weights = weights,
> start = start, etastart = etastart,
> --
> With Kind Regards,
> Zhi Jie,Zhang ,PHD
> Department of Epidemiology
> School of Public Health
> Fudan University
>


-- 
With Kind Regards,
Zhi Jie,Zhang ,PHD
Department of Epidemiology
School of Public Health
Fudan University

	[[alternative HTML version deleted]]

In this way you don't simulate response data; it'd also be a good idea 
to consider a bigger sample, e.g.,

dat <- matrix(rnorm(4000), ncol = 8)
dat <- data.frame(dat)
names(dat) <- paste("x", 1:8, sep = "")
dat$y <- rbinom(nrow(dat), 1, plogis(rowSums(dat)))

fit <- glm(y ~ ., family = binomial, data = dat)
fit

I hope it helps.

Best,
Dimitris

----
Dimitris Rizopoulos
Ph.D. Student
Biostatistical Centre
School of Public Health
Catholic University of Leuven

Address: Kapucijnenvoer 35, Leuven, Belgium
Tel: +32/(0)16/336899
Fax: +32/(0)16/337015
Web: http://med.kuleuven.be/biostat/
     http://www.student.kuleuven.be/~m0390867/dimitris.htm


----- Original Message ----- 
From: "zhijie zhang" <epistat at gmail.com>
To: <R-help at stat.math.ethz.ch>
Sent: Thursday, August 31, 2006 4:12 PM
Subject: [R] what's wrong with my simulation programs on logistic 
regression

> Dear friends,
> I'm doing a simulation on logistic regression model, but the 
> programs can't
> work well,please help me to correct it and give some suggestions.
> My programs:
> data<-matrix(rnorm(400),ncol=8)  #sample size is 50
> data<-data.frame(data)
> names(data)<-c(paste("x",1:8,sep=""))  #8
independent
> variables,x1-x8;
> #logistic regression model is logit(y)=x1+x2+x3+x4+x5+x6+x7+x8
>
data$y<-exp(data$x1+data$x2+data$x3+data$x4+data$x5+data$x6+data$x7+data$x8)/(1+(data$x1+data$x2+data$x3+data$x4+data$x5+data$x6+data$x7+data$x8))
>
> logist<-glm(y~.,family=binomial(),data=simdata)
> *Warning messages:*
> 1: algorithm can't converge in: glm.fit(x = X, y = Y, weights = 
> weights,
> start = start, etastart = etastart,
> 2: the probability is 0 or 1 in: glm.fit (x = X, y = Y, weights = 
> weights,
> start = start, etastart = etastart,
> -- 
> With Kind Regards,
> Zhi Jie,Zhang ,PHD
> Department of Epidemiology
> School of Public Health
> Fudan University
>
> [[alternative HTML version deleted]]
>
> ______________________________________________
> R-help at stat.math.ethz.ch 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.
> 

Disclaimer: http://www.kuleuven.be/cwis/email_disclaimer.htm

On Thu, 31 Aug 2006, zhijie zhang wrote:
> Dear friends,
>  I'm doing a simulation on logistic regression model, but the programs
can't
> work well,please help me to correct it and give some suggestions.
> My programs:
> data<-matrix(rnorm(400),ncol=8)  #sample size is 50
> data<-data.frame(data)
> names(data)<-c(paste("x",1:8,sep=""))  #8
independent variables,x1-x8;
> #logistic regression model is logit(y)=x1+x2+x3+x4+x5+x6+x7+x8
Rather it is logit(p) = ...,  and y ~ binomial(1, p)

There is a different sort of 'logistic regression' with 

y = exp(eta)/(1+exp(eta)) + epsilon

but you fit that by nls, not glm.
>
data$y<-exp(data$x1+data$x2+data$x3+data$x4+data$x5+data$x6+data$x7+data$x8)/(1+(data$x1+data$x2+data$x3+data$x4+data$x5+data$x6+data$x7+data$x8))
You need exp()/(1+exp()), and the second exp is missing.

Once you have p, you can use data$y <- rbinom(length(p), 1, p)
> logist<-glm(y~.,family=binomial(),data=simdata)
> *Warning messages:*
> 1: algorithm can't converge in: glm.fit(x = X, y = Y, weights =
weights,
> start = start, etastart = etastart,
> 2: the probability is 0 or 1 in: glm.fit (x = X, y = Y, weights = weights,
> start = start, etastart = etastart,
You do not have a Bernoulli response: it often helps to look at your 
simulated data to see if it makes sense (just as you would look at real 
data, I hope).

-- 
Brian D. Ripley,                  ripley at 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