I'm not sure I got the question, but I followed Crawley, Statistical
Computing page 413-415 for a similar problem last week. Hope it helps
Y~a-b*exp(-c*x)
plot your data "plot(Y,x)" and add some approx starting values.
a=asymptote of Y
b=a-intercept
c=curvature
#model:
exp.curve<- nls(Y~a-b*exp(-c*x), start=list(a=170,b=30,c=0.2))
#to plot and get trend line
plot(Y,x)
xaxis<-seq(0.7,7,0.01) "frm 0.7 to 7 in steps of 0.1
lines(xaxis,predict(exp.curve,list(x=xaxis)))
/Andreas
--------------------------
Dear Subscribers,
I tried to form a nonlinear model for :y~a+b*exp(x) where a & b are
specified.this expression is stated in the model field, but the message
I received was set_up nonlinear.How can this expression be stated with
s-plus 7?
thanks
lanre
|