I'll guess the model you want is as below, but use the curve fitting toolbox.
Your model might be defined as:
ft = fittype('sin((x - shift)/xscale)*yscale','coefficients',{'shift','xscale','yscale'})
ft =
General model:
ft(shift,xscale,yscale,x) = sin((x - shift)/xscale)*yscale
Now just call fit to fit the model to your data.
mdl = fit(X,Y,ft,'startpoint',[shiftguess,xscaleguess,yscaleguess]);
Other toolboxes have similar capability, but not quite as easy to use as the CFTB. (Stats TB, optimization TB)
