Hi Kenn,
Thanks for the suggestions, I'll have to see if I can figure out how to
convert the relatively simple call to lm with an equation and the data file
to the functions you mention (or if that's even feasible).
Not an expert in statistics myself, I am mostly concentrating on the
programming aspects of R. Problem is that I suspect my colleagues who
are providing some guidance with the stats end are not quite experts
themselves, and certainly new to R.
Cheers,
Esmail
Kenn Konstabel wrote:
lm does lots of computations, some of which you may never need. If speed
really matters, you might want to compute only those things you will
really use. If you only need coefficients, then using %*%, solve and
crossprod will be remarkably faster than lm
# repeating someone else's example
# lm(DAX~., EuStockMarkets)
y <- EuStockMarkets[,"DAX"]
x <- EuStockMarkets
x[,1]<-1
colnames(x)[1] <- "Intercept"
lm(y ~ x-1)
solve(crossprod(x), t(x))%*%y # probably this can be done more
efficiently
# and a naive timing
> system.time( for(i in 1:1000) lm(y ~ x-1))
user system elapsed
14.64 0.33 32.69
> system.time(for(i in 1:1000) solve(crossprod(x), crossprod(x,y)) )
user system elapsed
0.36 0.00 0.36
Also lsfit() is a bit quicker than lm or lm.fit.
Regards,
Kenn
______________________________________________
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.
