Regression (Cubic and Quadratic)
I am trying to run some regression analysis on a dataset I have but have very limited knowledge of Proc Reg. (i.e)
y = x
I have done the straight forward regaression, and have converted the variables of interest to the log(variable) and re ran the analysis.
log(y) = log(x)
I know want to do cubic and quadratic regression.
For cubic am I write in using the formula:
y = x x*x or should it just be y=x*x?
Additionally can anyone help me with how I shoud attempt the quadratic regression?
Sorry I am very much a novice at regression, so if I've made mistakes with my formulas above apologies.
Re: Regression (Cubic and Quadratic)
A quadratic regression would have 2nd-degree terms, and you should (at least initially) include all lower degree terms.
So the quadratic regression would be: y = x x*x
(ie what you called cubic in your post)
The cubic regression would be: y = x x*x x*x*x
Advertise on Talk Stats