Comment by abeppu
4 years ago
It is interesting to note that the Chatterjee paper makes a point of mentioning Pearson, Spearman, Kendall's tau, whereas the ones focused on in this paper all appear as citations but aren't explicitly discussed.
4 years ago
It is interesting to note that the Chatterjee paper makes a point of mentioning Pearson, Spearman, Kendall's tau, whereas the ones focused on in this paper all appear as citations but aren't explicitly discussed.
To be fair, Pearson, Spearman and Kendall's tau are the coeffisicent people use in practice. Had I been in the author's position I would have done the same: cite all the interesting developements but compare with what people actually use.
Comparing with something people barely know should be nice but people have a limited attention span so I would push to that anexes at best and focus on the more important parts.
For Pearson, Spearman, and Kendall's tau: (i) neither one of them could detect complex (nonlinear/non-monotone) dependence; (ii) they are clearly very powerful in detecting linear/monotone dependence. All these results have been documented in textbooks and repeatedly talked in classes for over 50 years.