univariate data

  • Formula:
  • Steps:
    • Estimate and from the data.
    • Compute the probability density for each data point.
    • Points with low probabilities (below a threshold ) are considered anomalies.

Multivariate Gaussian Distribution

Steps: - Extend the Gaussian model to include covariance across features. - Fit the multivariate Gaussian model:
- : Mean vector - : Covariance matrix - Threshold low-probability examples to identify anomalies.