[R] Regressão linear

Bernardo Rangel Tura tura at centroin.com.br
Fri Mar 6 09:56:29 CET 2009


On Thu, 2009-03-05 at 02:20 +0000, Ben Bolker wrote:
> Sueli Rodrigues <srodrigu <at> esalq.usp.br> writes:
> > 
> > Olá. Tenho um arquivo que a cada 6 linhas corresponde uma amostra da qual
> > preciso dos coeficientes da regressão linear. Como faço para que o
> > programa distinga a cada 6 linhas como uma amostra e não calcule como um
> > todo?
> > Estou usando a função: model=lm(y ~ x)
> > 
> 
>   You're more likely to get a response if you post to the list
> in English (even fractured English).
> 
>  Based on what Google translator thinks you said (you want
> to perform linear regressions on 6-line subsets of a data set?),
> here's a starting point (assuming your data are in a data frame
> mydata, and have column names x and y):
> 
> splitdat <- split(mydata,rep(1:6,length.out=nrow(mydata))
> linfits <- lapply(splitdata,lm,formula=y~x)
> coefs <- sapply(linfits,coef)
> 
> or something like that.
> 
>   Ben Bolker

Hi Ben Bolker 

First of all I would like to thank the kindness with my countrywoman.

Second in her problem each 6 rows is a subset for a linear regression so
the command is 

splitdat <- split(mydata,rep(1:(nrow(mydata)/6),each=6))

-- 
Bernardo Rangel Tura, M.D,MPH,Ph.D
National Institute of Cardiology
Brazil




More information about the R-help mailing list