

Nonlinear regression is an iterative procedure. The program must start with estimated initial values for each parameter. It then adjusts these initial values to improve the fit.
Every equation builtin to Prism, as well as equations you define, include rules to compute initial values. These rules use the range of the X and Y values to come up with initial values, which become the original automatic initial values. You can change the rules for userdefined equations, and can clone builtin equations to make them userdefined. The new rules will be invoked when you next choose this equation for a new analysis. It won't change the initial values for the analysis you are working on.
If nonlinear regression doesn't report any results, or if they seem wrong, the problem might be bad initial values. To check whether the initial values are reasonable, check "Don't fit the curve. Instead plot the curve defined by initial values." on the Diagnostics tab. When you click OK from the nonlinear regression dialog, Prism will not fit a curve but will instead generate a curve based on your initial values. If this curve is not generally in the vicinity of the data points, change the initial values before running nonlinear regression.
In most cases, the automatic initial values will work fine. If you have trouble getting a sensible fit, you might need to tweak the initial values. You can override the automatic initial values for each parameter for each data set:
1.Select a data set on the top part of the initial values tab. Or hold down the Ctrl key and select several data sets. Or click "Select all" in the top right of the tab.
2.Uncheck "choose automatically" and enter the new initial value.
you will find it easy to estimate initial values if you have looked at a graph of the data, understand the model, and understand the meaning of all the parameters in the equation. Remember that you just need an estimate. It doesn't have to be very accurate.
When fitting a polynomial model, it makes no difference what values you enter as initial values. When fitting other models, the importance of the initial values depends on how well the data define the curve and on how many parameters are in the model. Initial values matter the most when your data have a lot of scatter and don't define the model very well, and your model has many parameters.