Robust Covariance
Estimate robust covariance and detect multivariate outliers in JavaScript or TypeScript with MinCovDet and EllipticEnvelope from @kanaries/ml.
View as MarkdownRobust covariance
Robust covariance estimators describe the center and spread of multivariate data without letting a small number of extreme observations dominate the result. They are useful for anomaly detection, robust distance calculations, and preprocessing noisy tabular data.
@kanaries/ml provides browser- and Node.js-ready TypeScript implementations of FAST-MCD through MinCovDet and its anomaly-detection wrapper EllipticEnvelope.
Choose MinCovDet when you need robust location, covariance, support masks, or Mahalanobis distances. Choose EllipticEnvelope when you need predict, scoreSamples, and decisionFunction with a contamination threshold.
For classical, shrinkage, and sparse inverse covariance, see Covariance Estimators.
XGBoost Regressor
Learn what XGBoost Regressor does, when to use it, and how to run XGBoostRegressor in JavaScript or TypeScript with @kanaries/ml for browser and Node.js applications.
MinCovDet (FAST-MCD)
Compute robust multivariate location, covariance, support masks, and Mahalanobis distances with the MinCovDet JavaScript and TypeScript implementation in @kanaries/ml.