The Gaussian Maximum Likelihood Estimator (MLE) is able to spot outliers by computing a probability density function (PDF) over the features assuming they are independently and normally (Gaussian) distributed. Assigning low probability density to a sample translates to a high anomaly score.
Data Type Compatibility: Continuous
|1||contamination||0.1||float||The proportion of outliers that are assumed to be present in the training set.|
Return the column means computed from the training set:
public means() : array
Return the column variances computed from the training set:
public variances() : array
use Rubix\ML\AnomalyDetectors\GaussianMLE; $estimator = new GaussianMLE(6.0, 0.1);
- T. F. Chan et al. (1979). Updating Formulae and a Pairwise Algorithm for Computing Sample Variances.