R help - May 2004 - need help: regression models

hello,

I am working with some data which i would like to find the best fit model for
them.
First, this data are in a matrix. Each of the columns are treatments and the
raws are repetions for the each tratment.
So, i've already maden one chart with all the boxplot graphic, for each
treatment.
Now, i need to find the best fit model for them, i know that exponetial or log
model would be perfect in this situation (both can help me)

Therefore, i need to find the best exponential model for this data, and i would
like draw this tendence line in my boxplot (if it is possible, of course.)


if someone could help me i would be appreciated.
i am trying to solve these, but i am not having sucess.

best wishes for all.
Rodrigo Gazaffi




---




	[[alternative HTML version deleted]]

"Rodrigo" <rgazaffi at uol.com.br> said:
> I am working with some data which i would like to find the best fit model
> for them.
> First, this data are in a matrix. Each of the columns are treatments and
> the raws are repetions for the each tratment.
Is the response in a separate vector, or is it one of the columns?

...
> Now, i need to find the best fit model for them, i know that exponetial or
> log model would be perfect in this situation (both can help me)
Exponential models:
library(survival)
?Surv
?survreg

is the first thing I think of.

Log-linear models:
?glm

For either:
?predict
?fitted

to get the values to plot on the boxplot.

Hope that helps

Cheers

Jason