Struct StructuralModel 

pub struct StructuralModel {
    pub trend: TrendComponent,
    pub seasonal: Option<SeasonalComponent>,
    pub irregular_var: f64,
}

Basic Structural Model (BSM) combining trend, seasonality, and irregular.

Fields

trend: TrendComponent

Trend component (level + optional slope)

seasonal: Option<SeasonalComponent>

Optional seasonal component

irregular_var: f64

Irregular (observation noise) variance σ²_ε

Implementations

impl StructuralModel

pub fn new(period: Option<usize>) -> Result<Self>

Create a new structural model.

If period is Some(s), a seasonal component is included.

pub fn local_level(level_var: f64, obs_var: f64) -> Self

Build a local-level model (no slope, no seasonality).

pub fn local_linear_trend(level_var: f64, slope_var: f64, obs_var: f64) -> Self

Build a local linear trend model (with slope, no seasonality).

pub fn state_dim(&self) -> usize

Total state dimension.

pub fn to_ssm(&self) -> LinearGaussianSSM

Convert to a LinearGaussianSSM in state-space form.

State vector layout: [μ, (β), γ₁, γ₂, …, γ_{s-1}]

pub fn fit(&mut self, data: &[f64]) -> Result<f64>

Fit the model by maximising the Kalman filter log-likelihood.

Uses a simple coordinate-ascent / line-search optimiser on log-scale variance parameters (no external optimiser dependency). Returns the maximised log-likelihood.

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

Decompose the series into trend, seasonal, and irregular components.

Returns (trend, seasonal, irregular), each of length T.

Trait Implementations

impl Clone for StructuralModel

fn clone(&self) -> StructuralModel

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for StructuralModel

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.