Struct PcaAnomalyDetector 

pub struct PcaAnomalyDetector { /* private fields */ }

PCA-based multivariate anomaly detection.

Fits principal components retaining variance_threshold of total variance, then flags points with high reconstruction error.

Implementations

impl PcaAnomalyDetector

pub fn new( variance_threshold: Option<f64>, anomaly_threshold: Option<f64>, ) -> Self

Creates a new PCA anomaly detector with optional thresholds.

Trait Implementations

impl MultivariateAnomalyDetector for PcaAnomalyDetector

fn fit(&mut self, ctx: &MultiSeriesContext) -> Result<(), MultivariateError>

Fit the detector to the multi-series context.

fn detect( &self, ctx: &MultiSeriesContext, ) -> Result<Vec<MultivariateAnomalyScore>, MultivariateError>

Detect anomalies and return per-timestamp scores.