I would like to produce the R^square goodness-of-fit statistics for a predictive model.

I have the base data (10, 000 number of x-values) which are the true values given by an analytic/deterministic function x = function(a, b, c....). I have the prediction data (10,000 number of y-values) which are the prediction values from an engine.

Hence, the fitted equation should be y = 1 * x (as the prediction should be the deterministic output, all things being equal).

How would I be able to calculate the r^square of y relative to x, assuming a forced fitted line of y = 1 * x?

I believe the r^square standard formulae = 1 - SS_res/SS_tot; where SS_res = summation(difference between y and x)^2, and SS_tot = summation(y - x_mean?)^2.


Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Browse other questions tagged or ask your own question.