(1) | \rho_{xy} = \frac{{\rm cov}(x,y)}{\sigma(x) \sigma(y)} |
where
\{ (x, y) \} = \{ (x_1, y_1), \, (x_2, y_2), \, x_3 ... (x_n, y_n) \} | dynamic properties on the same discrete argument ( i=1..N) | ||
| cross-property covariance | ||
| |||
| standard deviation of property y |
Pearson correlation coefficient ranges between -1 and 1 and indicates how close the two properties can be related by a linear correlation:
y_i = a x_i + b, \quad \forall \, i=1..N |
with a certain pick on a and b.
Fig. 1. Highly correlated properties | Fig. 2. Poorely correlated properties | Fig. 3. Highly anti-correlated properties |