search for: regressioncoefficiants

...ogis. The regression works if I use a non selfstarting model. The formula is the same. But this needs very detailed information of the starting list. I dont have this Information all the time. An easy solution for selfstart model could be to add 1E-100 to the y==0 values. That works fine, but the Regressioncoefficiants differ to the original ones a bit. Sometimes this is problematic. So I wonder if it is possible to copy the code of SSlogis and rewrite the failing passage. I found this passage, but I dont have any Idea, where and how to create a new "SSlogis2". I guess this part of one SSlogis functi...