05-04-2017 05:19 AM
I am trying to find a functino that find the best-fit function according to some data points that I have. I am doing that with proc nlin, and have succeeded at computing it with a logarithmic model function. Currently I am trying to do it with a different model function* but I haven't managed to make it work.
I think I may be inputting the function with the wrong expression.
* The model function is the Weibull distribution whose expression is f(t) = 1 - e^(-(t/a)^b);
PROC NLIN DATA=aux_&s MAXITER=500 CONVERGE=1E-05 SINGULAR=1E-08 MAXSUBIT=30 OUTEST=coeficientes_segmento&s
NOITPRINT ; model segmento11 = (1 - (EXP(-(t/a)**b))); *Weibull distribution function; ; PARMS a=0.4 b=0.5 ; RUN;