headers
ANJAN PURKAYASTHA | 1 Feb 2009 02:05
Picon

question on statistical test

hi,
this is more of a statistical methodology question than an R question.
however, since there are quite a few expert statisticians in this forum i
think i can expect to get some useful feedback.
here is my problem:
i am studying a set of genes.
for each gene i have a set of probes (there are 2-5 probes per gene).
for each probe i have a set of 3 measurements (under conditions a, b and c).

let us say that for a given gene i have 3 probes X, Y and Z. and for each
probe there are 3 measurements, thus:

       a     b     c
X   Xa   Xb   Xc
Y   Ya   Yb    Yc
Z    Za   Zb    Zc

for each gene i need to test the correlation among the measurements under
each condition. so, are Xa, Ya, Za correlated AND Xb, Yb, Zb correlated AND
Xc, Yc, Zc correlated?
based on the degree of correlation among these probe measurements i will
choose the genes for further study.
of course i have to do this for each gene in my set of 100.

what statistical tests do i need to perform to accomplish this?

all advice will be appreciated.
tia,
anjan
(Continue reading)

Permalink | Reply | 0 comments |
headers
Christian Hoffmann | 1 Feb 2009 02:22
Picon

Simulation of linear lists

Hi there
has anybody experimented with the implementation in R of linear lists, 
as described e.g. in books on Pascal (Jensen and Wirth), or Wirth, 
"Algorithms and data structures" ? See 
\url{www.dbnet.ece.ntua.gr/~adamo/csbooksonline/AD.pdf }

Pointers would be welcome.

Christian

-- 

--

-- 
Christian W. Hoffmann,
Rigiblickstrasse 15 b, CH-8915Hausen am Albis, Switzerland,
Tel +41-44-7640853, c-w.hoffmann <at> sunrise.ch

______________________________________________
R-help <at> 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.
Permalink | Reply | 0 comments |
headers
Hesen Peng | 1 Feb 2009 02:29
Picon

Re: This site may harm your computer - Google warning about cran website

Hey all,

Just now I ran into this update on Google Blog explaining the whole incident:

http://googleblog.blogspot.com/2009/01/this-site-may-harm-your-computer-on.html

On Sat, Jan 31, 2009 at 11:03 AM, Duncan Murdoch <murdoch <at> stats.uwo.ca> wrote:
> On 31/01/2009 10:19 AM, Christos Argyropoulos wrote:
>>
>> This is extremely annoying (and it smells of rat). As of this time (1013
>> am ET), Google issues the following nasty warning: "This site may harm your
>> computer"
>> when I try to visit any of the CRAN package repository sites or CRAN task
>> view.
>> The funny thing is that Google's Safe Browsing Diagnostic page (it gives
>> diagnostics about any page that
>> it may contain Malware) does not seem to be working at the moment, so that
>> the problems that it found
>> with CRAN are not available for review.
>
> I suspect that the server that maintained those records was down, and their
> fallback was to assume that if it didn't respond, there was a problem.  The
> error message therefore showed up almost everywhere, but is fixed now.
>
> One speculation:  The error page told you you could go to the selected URL
> "at your own risk".  Does that mean they'll assume responsibility that all
> links are safe if they don't warn you?
>
> Duncan Murdoch
>
(Continue reading)

Permalink | Reply | 0 comments |
headers
Tom La Bone | 1 Feb 2009 03:13

Extracting Coefficients and Such from mle2 Output


The mle2 function (bbmle library) gives an example something like the
following in its help page. How do I access the coefficients, standard
errors, etc in the summary of "a"?

> x <- 0:10
> y <- c(26, 17, 13, 12, 20, 5, 9, 8, 5, 4, 8)
> LL <- function(ymax=15, xhalf=6)
+   -sum(stats::dpois(y, lambda=ymax/(1+x/xhalf), log=TRUE))
> a <- mle2(LL, fixed=list(xhalf=6))
> summary(a)
Maximum likelihood estimation

Call:
mle2(minuslogl = LL, fixed = list(xhalf = 6))

Coefficients:
     Estimate Std. Error z value     Pr(z)    
ymax  19.2881     1.7115  11.269 < 2.2e-16 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1 

-2 log L: 60.39244 

Tom

--

-- 
View this message in context: http://www.nabble.com/Extracting-Coefficients-and-Such-from-mle2-Output-tp21770860p21770860.html
Sent from the R help mailing list archive at Nabble.com.
(Continue reading)

Permalink | Reply | 8 comments |
headers
David Winsemius | 1 Feb 2009 05:11
Picon

Re: Extracting Coefficients and Such from mle2 Output


On Jan 31, 2009, at 9:13 PM, Tom La Bone wrote:

>
> The mle2 function (bbmle library) gives an example something like the
> following in its help page. How do I access the coefficients, standard
> errors, etc in the summary of "a"?

?coef
?vcov

eeep. Further comment on "etc" not possible at this time. Mind reading  
equipment malfunction.

--

-- 
David Winsemius

>
>
>> x <- 0:10
>> y <- c(26, 17, 13, 12, 20, 5, 9, 8, 5, 4, 8)
>> LL <- function(ymax=15, xhalf=6)
> +   -sum(stats::dpois(y, lambda=ymax/(1+x/xhalf), log=TRUE))
>> a <- mle2(LL, fixed=list(xhalf=6))
>> summary(a)
> Maximum likelihood estimation
>
> Call:
> mle2(minuslogl = LL, fixed = list(xhalf = 6))
>
(Continue reading)

Permalink | Reply | 6 comments |
headers
Bernardo Rangel Tura | 1 Feb 2009 09:28
Picon
Favicon

Re: How to analyse and model 2 time series, when one series needs to be differenced?

On Mon, 2009-01-26 at 08:52 +0000, Andreas Klein wrote:
> Hello.
> 
> How can I analyse the cross-correlation between two time series with ccf, if one of the time series need to
be differenced, so it is stationary?
> The two time series differ when in length and maybe ccf produces not the correct cross-correlation?!
> 
> Another problem:
> How can I model the two time series as an VARI-process with the dse package? - 
> So how can I handle it, that one series has to be differenced and the other series not?
> 
> I hope you can give me some hints.
> 
> 
> Regards,
> Andreas.

Hi Andreas,

If I understand your problem this script solve tour question

t<-1:15
x<-rnorm(10)
y<-.2-.3*t+rnorm(15)
y.dif<-diff(y,1)
ccf(x,y.dif)

--

-- 
Bernardo Rangel Tura, M.D,MPH,Ph.D
National Institute of Cardiology
(Continue reading)

Permalink | Reply | 0 comments |
headers
Veerappa Chetty | 1 Feb 2009 01:52
Picon

cluster function to use with GLM

Hi,In SPlus, I use the function 'cluster' with GLM to adjust for correlated
errors within a cluster ( e.g. cluster(ID)) as a covariate. Is there a
similar function in R to use with GLM ? I am new to R.  Thanks.

V.K. Chetty

--

-- 
Professor of Family Medicine
Boston University
Tel: 617-414-6221, Fax:617-414-3345
emails: chettyvk <at> gmail.com,vchetty <at> bu.edu

	[[alternative HTML version deleted]]

______________________________________________
R-help <at> 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.
Permalink | Reply | 1 comment |
headers
Brendan Cusick | 1 Feb 2009 06:07
Picon

Histogram height

Hello.  I have what I'm sure is a very simple question.  Hopefully someone
can help.  I can't seem to find the maximum height of a histogram.  In
matlab it would be simple to say max(hist(x, 100)), but this doesn't seem to
work in R.  Any input would be greatly appreciated.

Brendan

	[[alternative HTML version deleted]]

______________________________________________
R-help <at> 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.
Permalink | Reply | 2 comments |
headers
Prof Brian Ripley | 1 Feb 2009 09:53
Picon
Picon
Favicon

Re: cluster function to use with GLM

I am surprised: a Generalized Linear Model does not have clusters, nor 
did glm() allow clusters in any version of S-PLUS I have used.

I suspect this is a GEE model, for which see packages gee, geepack and 
yags (not on CRAN) under R.  But you can also have Generalized Linear 
Mixed Models, and there are several R packages to handle those.

On Sat, 31 Jan 2009, Veerappa Chetty wrote:

> Hi,In SPlus, I use the function 'cluster' with GLM to adjust for correlated
> errors within a cluster ( e.g. cluster(ID)) as a covariate. Is there a
> similar function in R to use with GLM ? I am new to R.  Thanks.
>
> V.K. Chetty
>
> -- 
> Professor of Family Medicine
> Boston University
> Tel: 617-414-6221, Fax:617-414-3345
> emails: chettyvk <at> gmail.com,vchetty <at> bu.edu
>
> 	[[alternative HTML version deleted]]
>
> ______________________________________________
> R-help <at> 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.
>
(Continue reading)

Permalink | Reply | 0 comments |
headers
Jim Lemon | 1 Feb 2009 11:44
Picon
Gravatar

Re: Histogram height

Brendan Cusick wrote:
> Hello.  I have what I'm sure is a very simple question.  Hopefully someone
> can help.  I can't seem to find the maximum height of a histogram.  In
> matlab it would be simple to say max(hist(x, 100)), but this doesn't seem to
> work in R.  Any input would be greatly appreciated.
>   
Hi Brendan,
Try:
max(hist(x,100)$counts,na.rm=TRUE)

Jim

______________________________________________
R-help <at> 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.
Permalink | Reply | 1 comment |

Gmane