| 1 | //! Codependence measures between two series: correlation-based distances, distance |
| 2 | //! correlation, mutual information and variation of information. |
| 3 | //! |
| 4 | //! These follow López de Prado, *Machine Learning for Asset Managers* (2020), chapter 3, and |
| 5 | //! are the distance layer under the hierarchical methods of AFML chapter 16 (see |
| 6 | //! [`crate::hrp`], [`crate::hcaa`] and [`crate::onc`]). |
| 7 | //! |
| 8 | //! - [`angular_distance`], [`absolute_angular_distance`] and [`squared_angular_distance`] |
| 9 | //! turn a Pearson correlation `rho` into a metric on `[0, 1]`. The first treats |
| 10 | //! `rho = -1` as maximally distant (long-only books); the other two as identical |
| 11 | //! (long-short books). |
nothing calls this directly
no outgoing calls
no test coverage detected