1

After some research, I find that the relationship between $R^2$ and the F distribution is as follows:

$$ R^2 = 1 - (1 + F \cdot \frac{p-1}{n-p})^{-1} $$

But I am not sure what the $p$ and $n$ values should be. Suppose my $R^2$ is calculated from 10 values of $x$ and $y$, what would $p$ and $n$ be?

Thanks!

EDIT: don't think it's a duplicate because my question is more basic, i.e. what $p$ and $n$ values are.

xyy
  • 411
  • 1
  • 5
  • 7
  • p is the number of parameters in model – TPArrow Jul 17 '15 at 22:21
  • @Hamed thanks for the answer, can you explain what parameters are? for example in a linear regression with one variable, are there two paramters (intercept and slope)? also is n the number of observations then? – xyy Jul 17 '15 at 23:24
  • yes, for example if you have a model like $Y=b_0+b_1X_1+\ldots+b_kX_k$ then $p=k+1$. Yes n is the number of observations in response, y. – TPArrow Jul 17 '15 at 23:37
  • You were right that it wasn't really a duplicate. I clarified one of the answers in the linked "duplicate" to define the terms $n$ and $p$ (that should have been defined there). – Glen_b Jul 18 '15 at 03:19

0 Answers0