Nonlinear regression is an iterative procedure. The program must start with estimated values for each variable that are in the right "ball park" - say within a factor of five of the actual value. It then adjusts these initial values to improve the fit.   

Prism automatically provides initial values for each variable, calculated from the range of your data. If you select a built-in equation, the rules for calculating the initial values are built-in to the program. If you enter a user-defined equation, you define the rules. See User-defined equations.

You'll 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. If you are having problems estimating initial values, set aside your data and simulate curves using the model. Change the variables one at a time, and see how they influence the shape of the curve. Once you have a better feel for how the parameters influence the curve, you might find it easier to estimate initial values.

To view and change the initial values:

1. Press the Initial values button on the Nonlinear Regression Parameters dialog.

2. Select a data set from the drop down list.

3. To change the initial value of a variable, deselect the Auto check box next to a variable and enter the new initial value.

After you fit data, Prism will use the best-fit values from that fit as the initial values for the next fit. If the changes you make to the data are minor, the fit will usually go much faster this way. But if you replace the data with values that are very different, the previous best-fit values may make very bad initial values, so bad that the nonlinear regression process may not converge on best-fit values. In this case, uncheck the option box "Use the results of the previous fit as initial values".

How much difference do initial values make?

When fitting a simple model to clean data, it won't matter much if the initial values are fairly far from the correct values. You'll get the same best-fit curve no matter what initial values you use, unless the initial values are very far from correct. Initial values matter more when your data have a lot of scatter or your model has many variables.

Viewing the curve generated by the initial values

If you aren't sure whether the initial values are reasonable, check "Don't fit, fix all variables to their initial values"  on the initial values dialog. When you 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.