Mahalanobis Depth in Python.

mhd.py

# -*- coding: utf-8 -*-
"""Mahalanobis Depth."""
# Author: Akira Tamamori <tamamori5917@gmail.com>
# License: BSD 2 clause

from sklearn.covariance import EmpiricalCovariance, MinCovDet
from sklearn.utils import check_array, check_random_state


class MHD:
    """MHD class.

    This class computes Mahalanobis depth.

    Parameters
    ----------
    cov_type: str, optional (default="emp")
        Covariance estimator.
        If ``cov_type'' is "emp", EmpiricalCovariance() is used;
        If ``cov_type'' is "mincov", MinCovDet() is used.

    random_state : int, RandomState instance or None, optional (default=None)
        If int, random_state is the seed used by the random number generator;
        If RandomState instance, random_state is the random number generator;
        If None, the random number generator is the RandomState instance
        used by np.random.
    """

    def __init__(self, random_state=None):
        super().__init__()
        self.cov_type = "emp"
        self.random_state = check_random_state(random_state)
        self.depth_values = None
        self.robust_cov = None

    def fit(self, X):
        """Fit detector.

        Parameters
        ----------
        X : numpy array of shape (n_samples, n_features)
            The input samples.

        Returns
        -------
        self : object
            Fitted estimator.
        """
        if self.cov_type == "emp":
            self.robust_cov = EmpiricalCovariance().fit(X)
        elif self.cov_type == "mincov":
            self.robust_cov = MinCovDet(random_state=self.random_state).fit(X)
        else:
            raise ValueError(
                f"cov_type={self.cov_type} " f"must be between 'emp' or 'mincov'."
            )
        self.depth_values = 1 / (1 + self.robust_cov.mahalanobis(X))
        return self

    def depth_function(self, X):
        """Predict raw depth values of X using the fitted object.

        Parameters
        ----------
        X : numpy array of shape (n_samples, n_features)
            The training input samples. Sparse matrices are accepted only
            if they are supported by the base estimator.

        Returns
        -------
        depth_values : numpy array of shape (n_samples,)
            The depth value of the input samples.
        """
        X = check_array(X)
        depth_values = 1 / (1 + self.robust_cov.mahalanobis(X))
        return depth_values