Understand Spearman’s Correlation for Beginners – Deep Learning Tutorial

Spearman’s correlation coefficient is a statistical measure of the strength of a monotonic relationship between paired data.

As to two paired data \(R\) and \(S\). The count in them is \(n\). Spearman’s correlation coefficient is defined as:

Here

\[D^2 = (R_i-S_i)^2\]

You should notice: \(R\) and \(S\) are rank.

r_s is in:

-1 ≤ r_s ≤ 1

where | r_s| is

0 -0.19  “very weak”
0.20 -0.39  “weak”
0.40 – 0.59  “moderate”
0.60 – 0.79  “strong”
0.80 – 1.0  “very strong”

So the | r_s| value of pair (x, y) is closed to 1, whic means they are strong monotonicity.

The bigger of | r_s| the better.