MD

class frlearn.data_descriptors.MD(preprocessors=())

Implementation of the Mahalanobis Distance (MD) data descriptor [1]. Mahalanobis distance is the multivariate generalisation of distance to the mean in terms of σ, in a Gaussian distribution. This data descriptor simply assumes that the target class is normally distributed, and uses the pseudo-inverse of its covariance matrix to transform a vector with deviations from the mean in each dimension into a single distance value. Squared Mahalanobis distance is χ²-distributed, the corresponding p-value is the confidence score.

Parameters

preprocessorsiterable = ()

Preprocessors to apply.

References

1

Mahalanobis PC (1936). On the generalized distance in statistics. Proceedings of the National Institute of Sciences of India, vol 2, no 1, pp 49–55.

class Model

Examples using frlearn.data_descriptors.MD

One class classification