Dear all,
I have a question on the four parameter logistic regression.
Not sure about how to set the start values for the parameters.
Can anyone familiar with this give some hint? Thanks so much.
Hannah
For example, suppose I have the following data values,
> temp
mean_std
[1,] 0.0 139
[2,] 0.5 142
[3,] 1.0 143
[4,] 2.0 145
[5,] 4.0 155
[6,] 7.0 169
[7,] 15.0 184
[8,] 30.0 220
[9,] 60.0 242
[10,] 150.0 269
[11,] 300.0 283
[12,] 600.0 289
[13,] 1500.0 291
[14,] 3000.0 299
[15,] 10000.0 299
logitfit <- nls(log(mean_std) ~ A+(B-A)/(1+(std[,1]/C)^scal), data=temp,
start=list(A=3.5, B=15, C=600, scale=2.5))
These starting values are not working!
[[alternative HTML version deleted]]
On Wed, Aug 15, 2012 at 9:08 PM, li li <hannah.hlx at gmail.com> wrote:> Dear all, > I have a question on the four parameter logistic regression. > Not sure about how to set the start values for the parameters. > Can anyone familiar with this give some hint? Thanks so much. > >See ?SSfpl which is the self-starter four parameter logistic, i.e. no starting values needed when you use it as in the example at the bottom of the help page. -- Statistics & Software Consulting GKX Group, GKX Associates Inc. tel: 1-877-GKX-GROUP email: ggrothendieck at gmail.com