Regression and Orthogonalisation

[minks]

Hi there,

I have to write code to perform the following task:

1. Run a simple regression with one independent variable; (e.g. Regress Y on X1)

2. Next I wish to introduce another independent variable in the model but only after removing its collinearity with the existing independent variable. (e.g. Introduce X2 after removing collinearity with X1)

3. To do the above, I want to use the orthogonalisation approach i.e. Regress the New Independent Variable on the Old Independent variable (Regress X2 on X1)

4. Create another series (say Z1) by summing up constant and residual of the above regression (Regression X2 on X1)

5. Introduce above series as independent variable in the original model (e.g. Y = C + b1 (X1) + b2 (Z1))

6. Do the above process for 300 variables...

I would greatly appreciate if any one could guide.

Regards,
Mayank