[R] Estimability of non-linear models
Vasco Cadavez
vcadavez at ipb.pt
Wed Jan 15 15:41:16 CET 2014
Dear R Users,
I'm fitting non-linear models using nlme package.
How can I check models estimability?
Data were obtained using an incomplete factorial design:
effect 1 = 4 levels
effect 2 = 5 levels
effect 3 = 5 levels
effect 4 = 5 levels
Cross table example for effect 1 and effect 2:
4 5 6 7 8
55 58 0 16 0 59
57 0 84 25 78 0
60 18 24 167 24 16
62 0 80 22 89 0
Cheers,
Vasco
On 14/01/14 11:00, r-help-request r-project.org wrote:
> Send R-help mailing list submissions to
> r-help r-project.org
>
> To subscribe or unsubscribe via the World Wide Web, visit
> https://stat.ethz.ch/mailman/listinfo/r-help
> or, via email, send a message with subject or body 'help' to
> r-help-request r-project.org
>
> You can reach the person managing the list at
> r-help-owner r-project.org
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of R-help digest..."
>
>
> Today's Topics:
>
> 1. To modify FMCD (Hock Ann Lim)
> 2. Re: computation (Lisa S)
> 3. Re: Find the package of a class given classname (Hadley Wickham)
> 4. Re. Confidence intervals for prediction with R2Bayesx (garth)
> 5. Re: HOW TO CREATE A SUBSET FROM A LARGE DATASET based on some
> conditions (arun)
> 6. Re: R riddle (arun)
> 7. tcltk makes R crash under Mavericks (Claudia Tebaldi)
> 8. Re: tcltk makes R crash under Mavericks (Berend Hasselman)
> 9. Re: tcltk makes R crash under Mavericks (Claudia Tebaldi)
> 10. Re: tcltk makes R crash under Mavericks (Berend Hasselman)
> 11. Re: tcltk makes R crash under Mavericks (John Fox)
> 12. Barplot not showing all labels (Jeff Johnson)
> 13. coxph with time-dependent covariates (Phil Stevenson)
> 14. Re: Barplot not showing all labels (Jim Lemon)
> 15. Re: coxph with time-dependent covariates (David Winsemius)
> 16. How to use ddply (Amitabh Dugar)
> 17. likelihood ratio test for mean difference assuming unequal
> variance (Amanda Li)
> 18. Re: How to use ddply (arun)
> 19. Re: SE for R-squared (peter dalgaard)
> 20. Re: How to use ddply (David Winsemius)
> 21. Re: Barplot not showing all labels (William Dunlap)
> 22. Plot only a portion of a record (Janet Choate)
> 23. Re: Plot only a portion of a record (Jim Lemon)
> 24. Re: Plot only a portion of a record (David Winsemius)
> 25. Re: Plot only a portion of a record (arun)
> 26. Re: likelihood ratio test for mean difference assuming
> unequal variance (Ben Bolker)
> 27. Running multidimensional regressions (andrews Nikolaiv)
> 28. "squashed" domains when doing PCA with bio3d (Bianca Ihrig)
> 29. Re: "squashed" domains when doing PCA with bio3d
> (Tsjerk Wassenaar)
> 30. Re: Running multidimensional regressions (David Winsemius)
> 31. How fast is the JSS reviewing process? (Christophe Genolini)
> 32. Re: How fast is the JSS reviewing process? (Achim Zeileis)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Mon, 13 Jan 2014 03:52:30 -0800 (PST)
> From: Hock Ann Lim <lim_ha yahoo.com>
> To: R-Help <r-help r-project.org>
> Subject: [R] To modify FMCD
> Message-ID:
> <1389613950.86125.YahooMailNeo web160501.mail.bf1.yahoo.com>
> Content-Type: text/plain
>
> Dear R Experts,
>
> Can anyone teach me how to modify the FMCD algorithm in R. If I wish to replace the C-step in FMCD by the index set.
> Let I(old)={pi(1)old,pi(2)old,...,p(h)old} and I(new)={pi(1)new, pi(2)new,..,p(h)new} the index sets that correspond to the sample items in H(old) and H(new) respectively.
>
> The C-step is conducted as follow:
> If I(new) not equal to I(old), H(old):=H(new). and calculate T and S based on H(new). Otherwise, the process is stopped.
>
> Thank you.
>
> Regards,
> Lim Hock Ann
>
> [[alternative HTML version deleted]]
>
>
>
> ------------------------------
>
> Message: 2
> Date: Mon, 13 Jan 2014 16:01:31 +0800
> From: Lisa S <lisha.sheng gmail.com>
> To: IZHAK shabsogh <ishaqbaba yahoo.com>
> Cc: "r-help r-project.org" <r-help r-project.org>
> Subject: Re: [R] computation
> Message-ID:
> <CAOsPdsvnbwt6gBKdo88fTYxpssfAeXA0GuwQu-9wNWe94YL7xw mail.gmail.com>
> Content-Type: text/plain
>
> In this calculation:
> (dg0%*%t(dx0))
> is
>
> [,1] [,2]
> [1,] 0 -0.75
> [2,] 0 0.50
>
> So: exp2 = NaN
>
> It is extremely easy to find out the issue if you go back line by line,
> this is a basic procedure of debugging. Please do it yourself next time.
>
>
>
>
> On Mon, Jan 13, 2014 at 3:09 PM, IZHAK shabsogh <ishaqbaba yahoo.com> wrote:
>
>> kindly help me fine what is the mistake with following:
>> my aim is to compute those steps and obtain a vector with values (3,5)
>> but i am geting (NAN,NAN)
>>
>> Q<-matrix(c(5,-3,-3,2),2,2)
>> b<-rbind(0,1)
>> H0<-diag(2)
>> x0<-rbind(0,0)
>> d0<-b
>> g0<--b
>> a0<--(t(g0)%*%d0)/(t(d0)%*%Q%*%d0)
>> x1<-x0+a0[,1]*d0
>> dx0<-x1-x0
>> g1<-Q%*%x1-b
>> dg0<-g1-g0
>> exp1<-(t(dg0)%*%H0%*%dg0)/(t(dg0)%*%dx0)
>> exp2<-(dx0%*%t(dx0))/(dg0%*%t(dx0))
>> exp3<-(dx0%*%t(dg0)%*%H0 +H0%*%dg0%*%t(dx0))/(dx0%*%t(dg0))
>> H1<-H0+(1+exp1[,1])*exp2-exp3
>> d1<--H1%*%g1
>> a2<-(t(g1)%*%d1)/(t(d1)%*%Q%*%d1)
>> x2<-x1+a2[,1]*d1
>>
>> [[alternative HTML version deleted]]
>>
>>
>> ______________________________________________
>> R-help r-project.org 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.
>>
>>
> [[alternative HTML version deleted]]
>
>
>
> ------------------------------
>
> Message: 3
> Date: Mon, 13 Jan 2014 08:42:09 -0600
> From: Hadley Wickham <h.wickham gmail.com>
> To: Herv? Pag?s <hpages fhcrc.org>
> Cc: R-help <r-help r-project.org>
> Subject: Re: [R] Find the package of a class given classname
> Message-ID:
> <CABdHhvFtxOZY6Ghfw4o=c_rH9dexMG7k8McHgjGPHGEroYcxgQ mail.gmail.com>
> Content-Type: text/plain; charset=UTF-8
>
>>> How to find the package of a class given classname?
>>> For example, there is a class called GAlignments, I want to do something
>>> like
>>> attr("GAlignments", "package") that gives you the package where the class
>>> is defined? But of course, attr("GAlignments", "package") won't work...
>> You didn't say whether you wanted to be able to do this
>> programmatically.
>>
>> If that's the case maybe you're lucky and the package where the class
>> is defined is already attached to your session. In that case:
>>
>> > attr(class(new("GAlignments")), "package")
>> [1] "GenomicRanges"
> Or maybe findClass?
>
>> Otherwise, if the package in question is installed but not attached
>> you can still do:
>>
>> ??`GAlignments-class`
> Or class?GAlignments
>
> Hadley
>
--
Vasco A. P. Cadavez, PhD
Departamento de Ciência Animal & Centro de Investigação de Montanha (CIMO)
Escola Superior Agrária do Instituto Politécnico de Bragança
Campus de Santa Apolónia, Apartado 1172
5301-854 Bragança Portugal
Telefone: (+351) 273 303 304
Fax: (+351) 273 325 405
e-mail: vcadavez ipb.pt
More information about the R-help
mailing list