[R] Fitting a linear model with a break point

Daniel Brewer daniel.brewer at icr.ac.uk
Tue Sep 8 16:15:43 CEST 2009


Hello,

I would like to test some data to see whether it has the shape of a step
function (i.e. y1 up until x_th and then y2 where x_th is the
threshold).  The threshold x_th is unknown and the x values can only
take discrete values (0,1,2,3,4).

An example would be:
data<- data.frame(x=1:20,y=c(rnorm(10),rnorm(10,10)))


I was thinking along the lines of fitting some sort of piiecewise linear
model which has the gradient constrained to zero trying out all possible
different threshold and taking the one with the least residuals.  I am
not sure how to implement this in R.  Anyone got any ideas?

Also is there a way of including the threshold in the actual model, so
that could be estimated too?

Thanks

Dan

-- 
**************************************************************
Daniel Brewer, Ph.D.

Institute of Cancer Research
Molecular Carcinogenesis
Email: daniel.brewer at icr.ac.uk
**************************************************************

The Institute of Cancer Research: Royal Cancer Hospital, a charitable Company Limited by Guarantee, Registered in England under Company No. 534147 with its Registered Office at 123 Old Brompton Road, London SW7 3RP.

This e-mail message is confidential and for use by the a...{{dropped:2}}




More information about the R-help mailing list