See the result first:
Approximated Val: 28.5041 Actual Val:14.1421
My approach is first calculate the linear regression function for each variable, no matter how they are co-related. I don't care at the moment. And it sums up the total of all intercepts and break them up into 4 slopes which contributes to each component (4 IVs), then the result is about the double of the actual value, why?
thanks
Jack