Greetings,
I would like to calculate tolerancelimits for a series of predicted values
from nonlinear regression models. I've been using the tolerance package
but the self-starting functions are not in the derivitive tables. When I
spellout the functions and supply starting values I repeatedly get an
error message regardless of the starting values I use (see output below).
Does anyone have a solution for this problem?
> oh_fit2<-nls(MBB ~ SSfpl(AGE, a, b, c, d))
> oh_fit2
Nonlinear regression model
model: MBB ~ SSfpl(AGE, a, b, c, d)
data: parent.frame()
a b c d
9.308 36.420 67.982 12.815
residual sum-of-squares: 46834
Number of iterations to convergence: 9
Achieved convergence tolerance: 8.727e-06 > tol<-nlregtol.int(oh_fit2,side = 2, alpha = 0.05, P = 0.95)
The following object(s) are masked from 'temp (position 3)':
a, b, c, d
Error in deriv.formula(form, beta.names) :
Function 'SSfpl' is not in the derivatives
table> oh_fit3<-nls(MBB ~ a + ((b - a)/(1 + exp((c-AGE)/d))), start = list(a =
10, b = 35, c = 65, d = 10))> oh_fit3
Nonlinear regression model
model: MBB ~ a + ((b - a)/(1 + exp((c - AGE)/d)))
data: parent.frame()
a b c d
9.307 36.421 67.982 12.815
residual sum-of-squares: 46834
Number of iterations to convergence: 9
Achieved convergence tolerance: 7.474e-06 > tol<-nlregtol.int(oh_fit3,side = 2, alpha = 0.05, P = 0.95)
Error: Error in nls routine. Consider different starting estimates
of the parameters. Type help(nls) for more options.
Thank you in advance!
Grant.
___________________________
Grant M. Domke
Forest Inventory and Analysis
U.S. Forest Service
Northern Research Station
1992 Folwell Ave.
St. Paul, MN 55108
Ph: 651.649.5138
Fax: 651.649.5140
[[alternative HTML version deleted]]