If you have a set of data (Xi,Yi) and you have a model
Yhat=f(Xhat) where f includes parameters alpha1, alpha2, etc.
For a given alpha1, alpha2, you can compute the residual (error) of the estimate as
Ri = Yi - Yhat(Xi)
The least squared estimate f uses values of alpha1, alpha2 etc that minimizes the sum of the squares of the residuals.
This is very easily implemented using the solver add-in of excel. You put in initial guesses for alpha1, alpha2, compute the Yhati and the Ri, create a cell with the sum of the squares of the residuals, and use solver to adjust the alpha's to minimize that sum of residuals.