Mail Archive Search: <+from:vitomuggeo@hotmail.com>

Results:

References: [ +from:vitomuggeo@hotmail.com: 20 ]

Total 20 documents matching your query.

1. problem in fitting model with vectors and data.frame (score: 1): Author: "vito muggeo" <vitomuggeo@hotmail.com>; Date: Sat, 01 Dec 2001 10:39:07; Hi all, this is (at least for me) a very strange problem (almost a puzzle). My data.frame has three variable y, x, f summary(mydata) x y f Min.:16.71 Min.: 0.000 Min.:1.000 ..... ...... ...... Max.:2; /archives/html/s-news/2001-12/msg00000.html (8,681 bytes)
2. warning in dropping global assign (score: 1): Author: "vito muggeo" <vitomuggeo@hotmail.com>; Date: Wed, 05 Dec 2001 08:38:11; Dear all, I've written some function where global assignements are requested, for instance: fn<-function(....){ x<<-.... y<<-.... .... rm(x,y, w=1) return(....)} Everything works but S+ with the outp; /archives/html/s-news/2001-12/msg00033.html (7,116 bytes)
3. Re: Std. Error of coefficients in GAM Model (score: 1): Author: "vito muggeo" <vitomuggeo@hotmail.com>; Date: Thu, 20 Dec 2001 13:54:06; I really don't understand what you mean. Multicollinearity is a mathematical problem, not statistical. If, as limit, X2=k*X1, the model matrix (X'X)^(-1) is not invertible, and the algorithm should; /archives/html/s-news/2001-12/msg00171.html (8,574 bytes)
4. predict() in models with sigma/mu constant (score: 1): Author: "vito muggeo" <vitomuggeo@hotmail.com>; Date: Tue, 13 Nov 2001 13:50:02; dear S-Plus users, I'm interested in fitting linear (or non-linear maybe) models with constant coefficient of variation. Something like x<-1:100 mu<-3+2*x y<-rnorm(100, mu, 2*mu). The model: obj<-glm; /archives/html/s-news/2001-11/msg00101.html (7,136 bytes)
5. compilated function (score: 1): Author: "vito muggeo" <vitomuggeo@hotmail.com>; Date: Tue, 30 Oct 2001 13:07:32; Dear all, I'm intersted in using some function from Terry Therneau available at http://www.mayo.edu/hsr/biostat.html Unfourtanately just source codes are available. Is there anyone that has compilate; /archives/html/s-news/2001-10/msg00203.html (6,788 bytes)
6. hat matrix for gam (score: 1): Author: @vip.cybercity.dk; Date: Thu, 30 Aug 2001 15:23:33; Hi all, this isn't a real S+ question . I apologize for this I'm interested in calculating hat matrix for GAM. The problem is the model.matrix doesn't depend on df of the smoothing term, and so model; /archives/html/s-news/2001-08/msg00237.html (6,821 bytes)
7. nlme() problem (score: 1): Author: "vito muggeo" <vitomuggeo@hotmail.com>; Date: Tue, 24 Jul 2001 20:36:23; Dear S+ users, Do you know what is the minimum value of observations for each Subj to fit a logistic model with nlme()? I've some problem in fitting the model: For istance, my (grouped) dataframe is; /archives/html/s-news/2001-07/msg00171.html (9,738 bytes)
8. second derivatives calculation (score: 1): Author: "vito muggeo" <vitomuggeo@hotmail.com>; Date: Thu, 26 Jul 2001 11:26:20; Hi all, I used the following code to get the value of myfn() and its first derivatives at x=10 with K=2 and U=3, for istance myfn<-deriv(~K*x+U*x/K, c("K","U"), function(x,K,U){}) K<-2; U<-3 myfn(10,; /archives/html/s-news/2001-07/msg00189.html (6,968 bytes)
9. strange profile likelihood result (score: 1): Author: "vito muggeo" <vitomuggeo@hotmail.com>; Date: Sat, 09 Jun 2001 06:25:07; Hi all, I have the following (rather strange) problem: I fit a gam model by means of obj<-gam(y~x1+s(x2)+...+Z,family=poisson) and used summary.glm(obj, corr=F) ... Coefficients: Value Std. Error t v; /archives/html/s-news/2001-06/msg00077.html (9,366 bytes)
10. Double random effect (score: 1): Author: "vito muggeo" <vitomuggeo@hotmail.com>; Date: Mon, 09 Apr 2001 21:10:48; Hi all, I would like to fit simple Poisson regression model for count data. However I've several clusters and furthermore I know that some important explanatory variable is missing. So I should fit a; /archives/html/s-news/2001-04/msg00106.html (7,604 bytes)
11. Re: Negative-binomial mixed effect models (score: 1): Author: "vito muggeo" <vitomuggeo@hotmail.com>; Date: Tue, 10 Apr 2001 19:51:48; Dear Prof. Ripley, thanks for your replay, reglm() is a S+ function to fit GLM with random effect using the method of Shall (1991). Look at: http://www.maths.uq.edu.au/~gks/s Do you think is it possi; /archives/html/s-news/2001-04/msg00116.html (6,920 bytes)
12. persp() function (score: 1): Author: "vito muggeo" <vitomuggeo@hotmail.com>; Date: Fri, 13 Apr 2001 13:57:20; Dear S+ users, I'm using persp() to get 3D plots: persp(outer(x,y,f)) or persp(interp(x, y, z)). The output is fine, but I cannot rotate the graphic. Probably it's silly, but I really can't understan; /archives/html/s-news/2001-04/msg00152.html (6,876 bytes)
13. [S] R: repetead measures using lme (score: 1): Author: "vito muggeo" <vitomuggeo@hotmail.com>; Date: Thu, 06 Jul 2000 07:32:27 PDT; Hi Janine, In general using random effect models to fit unbalanced longitudinal data (i.e., with different number of times) could be a right choice. You have to specify just the cluster variable=the; /archives/html/s-news/2000-07/msg00029.html (9,058 bytes)
14. Re :[S] Conditional Logistic Regression (score: 1): Author: "vito muggeo" <vitomuggeo@hotmail.com>; Date: Wed, 19 Jul 2000 16:51:12 GMT; Hi Stephen, The conditional Likelihood is egual to partial likelihood in Cox model when the method to handle ties is 'exact' (the risk set would represent the matched set in the logistic model and th; /archives/html/s-news/2000-07/msg00129.html (7,944 bytes)
15. [S] score test in GLM (score: 1): Author: "vito muggeo" <vitomuggeo@hotmail.com>; Date: Fri, 05 May 2000 04:36:49 PDT; Hi all, probably this is a silly question: I'm interested in perfoming the (efficient) score test for any subset of coefficient in glm() (or gam()). I looked at some books (Venables&Ripley,Selvin...); /archives/html/s-news/2000-05/msg00031.html (7,374 bytes)
16. [S] bootsrap and logistic model (score: 1): Author: "vito muggeo" <vitomuggeo@hotmail.com>; Date: Mon, 10 Apr 2000 09:05:15 PDT; Hi all, Let y1/n1, y2/n2 and y3/n3 three binomial samples that I would compare by means of any link for binary data (logit, probit or log-log). However if y1=0 ML estimates don't exist, so I should u; /archives/html/s-news/2000-04/msg00091.html (9,276 bytes)
17. Re:[S] Bootstrap and logistic model (score: 1): Author: "vito muggeo" <vitomuggeo@hotmail.com>; Date: Tue, 11 Apr 2000 01:52:10 PDT; Hi all, Thanks to B.Venables, B.D.Ripley, P.Ellis and R.Lancelot for their useful comment. Right syntax is: glm(y/n~x,family=binomial,weights=n). Does anyone know whether is possible to built bootstr; /archives/html/s-news/2000-04/msg00100.html (9,744 bytes)
18. Re: [S] Question about GAM Model Selection using AIC (score: 1): Author: "vito muggeo" <vitomuggeo@hotmail.com>; Date: Wed, 12 Apr 2000 00:55:08 PDT; I don't know your data, but I can tell you something about selecting df in each smoothed variable by means of AIC, i.e. by using step.gam. Sometime the AIC select too many df and so the almost the al; /archives/html/s-news/2000-04/msg00117.html (9,686 bytes)
19. [S] simulating contingency tables (score: 1): Author: vito muggeo <vitomuggeo@hotmail.com>; Date: Tue, 29 Jun 1999 02:48:43 PDT; I would like to simulate contingency tables in S-plus. Can anyone to suggest me the way to do it, and maybe some paper/book about the topic? Thanks to all in advance, vito. -- Vito Muggeo Istituto di; /archives/html/s-news/1999-06/msg00232.html (7,378 bytes)
20. [S] semiparametric proportional odds model in survival analysis (score: 1): Author: "vito muggeo" <vitomuggeo@hotmail.com>; Date: Tue, 27 Apr 1999 10:02:32 PDT; Hi to all, The very popular Cox's model is a semiparametric proportional hazard model for survival data: however the semiparametric proportional odds model could be a valid alternative (Bennett, 1983; /archives/html/s-news/1999-04/msg00251.html (7,374 bytes)

Current List: 1 - 20
Page: [1]