[R] result of rqss

[David Winsemius]

If you do a search on the quantreg documentation for "piecewise", the  
first hit on the pdf has code on the same page. Running that rqss fit  
call unaltered with your data produced a straight line (because the  
data only has a domain of 0-0.5) but lowering lambda lets the  
piecewise character get captured. Try:

fit <- rqss(y ~ qss(x, lambda = .05),tau = .9)
plot(fit)

On Jun 19, 2009, at 9:52 AM, Michael Hecht wrote:

> Hello,
>
> i have the following data:
>
> x 
> = 
> c 
> (0,0.02,0.03,0.04,0.05,0.06,0.07,0.08,0.09,0.1,0.11,0.12,0.13,0.14,0.15,0.16,0.17,0.18,0.19,0.2,0.21,0.22,0.23,0.25,0.26,0.27,0.46,0.47,0.48,0.49 
> )
> y 
> = 
> c 
> (0.48,0.46,0.41,0.36,0.32,0.35,0.48,0.47,0.55,0.56,0.54,0.67,0.61,0.60,0.54,0.51,0.45,0.42,0.44,0.46,0.41,0.43,0.43,0.48,0.48,0.47,0.39,0.37,0.32,0.29 
> )
>
> and tried to get piecewise linear regression. Doing a simple spline  
> smoothing gives the basic shape of the expected curve:
>
> plot(x,y)
> lines(smooth.spline(x, y), lty=2, col = "red")
>
> Now I tried to do rqss. But what I got was only a straight line and  
> nothing appropriate.
> No parameter change did help. Can anyone tell me what's wrong with  
> this approach?
>
> library(quantreg)
> fit <- rqss(y ~ qss(x))
> plot(fit)
> points(x,y)
> lines(smooth.spline(x, y), lty=2, col = "red")
>
> Thank You in advance.
> ______________________________________________________

David Winsemius, MD
Heritage Laboratories
West Hartford, CT