How do I do the regression with multi-independent variables?

[Guest]

I have about 2-3 independent variable with only one dependent variable. For
example. I have Z as function of X,Y. "Z=f(X,Y)" I want to do the regression
to get the equation that show how X, Y relate to Z. If anyone knows, please
contact me or post the answer on the web. I bet there are a lot of people out
there who want to know the answer also.

Thanks

Scott

 

[Jerry W. Lewis]

If the function is linear in the unknown coefficients, then use LINEST
(it is not nearly as limited as Help might lead you to believe).

If the function is nonlinear in the unkown coefficients then you will
have to work a little harder. Make some initial guesses for the unknown
coefficients and place those guesses in worksheet cells (one coefficient
per cell). In other cells, calculate predicted values (based on cell
references to the guessed coeficients) for each observed data point. In
another cell calculate the sum of squared deviations between observed
data and predicted values, for example
=SUMPRODUCT((dataRange-predictedRange)^2)
Then use Solver to minimize the sum of squared deviations by changing
the values in the coefficient guesses range.

Jerry