I am trying to fit a set of experimental data to a model with four parameters using NonlinerRegress. The problem is that the model is very complex i.e. it is a complicated double integral. The first integral can only be calculated numerically, and the second integral is calculated after interpolation of the table generated from the first integral. The result after the second integral is the "model" but I think Mathematica is unable to compute the first intergral because it "isn't numerical" as the value of the parameters are not specified until the program executes NonlinearRegress. Is there any way this problem can be resolved?