[R] lasso for variable selection

[zubin]

Attended JSM last week and Friedman mentioned the use of LASSO for 
variable selection (he uses it for rules ensembles).  I am an 
econometrician and not familiar with, i started running the examples in 
R this week and you get to the plots section of the LARS package.   
Plots of beta/max(beta)  vs standardized coefficients.  How does one 
interpret them?  u see plots of each variable converging to zero at 
different times - its pretty cool - but can i use this for variable 
importance?

for variable selection - i have a group of correlated variables that we 
need to determine importance in predicting change of a Y variable.

-zubin