0

I once heard the statement that 

Covariance is a measure of the linear relationship between two variables, such as X and Y

If X and Y has non-linear relationship, what will be the metric corresponding to Covariance

user3269
  • 4,622
  • 8
  • 43
  • 53
  • You can get some ideas here: https://stats.stackexchange.com/questions/65829/can-pca-be-extended-to-account-for-nonlinear-dependencies, https://stats.stackexchange.com/questions/218127/intuition-behind-pearson-correlation-co-variance-and-cosine-similarity/218158#218158 and https://stats.stackexchange.com/questions/17618/what-can-i-do-beyond-pearson-correlation/50680#50680 – kjetil b halvorsen Apr 16 '19 at 16:17
  • 1
    It depends partly how general you want to be. Do you want to measure the strength of monotonic relationships? General functional relationships? Relations that are *not* functions (like $X^2+Y^2 = r^2$)? Completely general forms of dependence? – Glen_b Apr 17 '19 at 01:20

0 Answers0