Struct Ewma 

pub struct Ewma { /* private fields */ }

EWMA chart parameters.

Implements the Exponentially Weighted Moving Average control chart for detecting small sustained shifts in the process mean.

Examples

use u_analytics::detection::Ewma;

let ewma = Ewma::new(50.0, 2.0).unwrap();
// In-control data
let data = [50.5, 49.8, 50.2, 49.9, 50.0, 50.1, 49.7, 50.3];
let results = ewma.analyze(&data);
assert_eq!(results.len(), data.len());
assert!(ewma.signal_points(&data).is_empty());

Reference

Roberts, S.W. (1959). “Control Chart Tests Based on Geometric Moving Averages”, Technometrics 1(3).

Implementations

impl Ewma

pub fn new(target: f64, sigma: f64) -> Option<Self>

Creates a new EWMA chart with the given target mean and standard deviation.

Uses default parameters lambda=0.2 and L=3.0.

Returns

None if sigma is not positive or finite, or if target is not finite.

Reference

Roberts (1959), Technometrics 1(3). Default lambda=0.2, L=3.0 provides good sensitivity for detecting shifts of 0.5-2.0 sigma.

pub fn with_params( target: f64, sigma: f64, lambda: f64, l_factor: f64, ) -> Option<Self>

Creates an EWMA chart with custom parameters.

Parameters

• target: Process target mean (mu_0, must be finite)
• sigma: Process standard deviation (must be positive and finite)
• lambda: Smoothing constant (must be in (0, 1])
• l_factor: Control limit width factor (must be positive and finite)

Returns

None if any parameter is invalid.

pub fn analyze(&self, data: &[f64]) -> Vec<EwmaResult>

Analyzes a sequence of observations and returns EWMA statistics for each point.

The EWMA statistic is initialized to the target mean (Z_0 = mu_0). Non-finite values in the data are skipped (the previous EWMA value is carried forward).

Examples

use u_analytics::detection::Ewma;

let ewma = Ewma::new(50.0, 2.0).unwrap();
// Data with shift
let mut data: Vec<f64> = vec![50.0; 10];
data.extend(vec![55.0; 10]); // large shift
let signals = ewma.signal_points(&data);
assert!(!signals.is_empty()); // shift detected

Complexity

Time: O(n), Space: O(n)

pub fn signal_points(&self, data: &[f64]) -> Vec<usize>

Returns the indices of observations where an EWMA signal occurred.

A signal occurs when the EWMA statistic exceeds the upper or lower control limit.

Complexity

Time: O(n), Space: O(k) where k is the number of signal points