Enum ModelParams 

pub enum ModelParams {
    Ses {
        alpha: f64,
        level: f64,
        residual_std: f64,
    },
    HoltLinear {
        alpha: f64,
        beta: f64,
        level: f64,
        trend: f64,
        phi: f64,
        residual_std: f64,
    },
    HoltWinters {
        alpha: f64,
        beta: f64,
        gamma: f64,
        period: usize,
        level: f64,
        trend: f64,
        seasonal: Vec<f64>,
        multiplicative: bool,
        residual_std: f64,
        seasonal_phase: usize,
    },
    Arima {
        p: usize,
        d: usize,
        q: usize,
        ar_coeffs: Vec<f64>,
        ma_coeffs: Vec<f64>,
        constant: f64,
        residual_std: f64,
        effective_ar_order: Option<usize>,
    },
    Sarima {

        p: usize,
        d: usize,
        q: usize,
        sp: usize,
        sd: usize,
        sq: usize,
        m: usize,
        ar_coeffs: Vec<f64>,
        ma_coeffs: Vec<f64>,
        sar_coeffs: Vec<f64>,
        sma_coeffs: Vec<f64>,
        constant: f64,
        residual_std: f64,
    
},
    LinearRegression {
        slope: f64,
        intercept: f64,
        r_squared: f64,
        residual_std: f64,
        start_ts: i64,
        interval_ns: i64,
    },
    Custom {
        name: String,
        data: Vec<u8>,
    },
}

Fitted model parameters — variant per model type.

Variants

Ses

Simple Exponential Smoothing parameters.

Fields

alpha: f64

Smoothing factor α ∈ (0, 1).

level: f64

Last smoothed level.

residual_std: f64

Standard deviation of in-sample residuals.

HoltLinear

Holt’s Linear Trend parameters.

Fields

alpha: f64

Level smoothing factor α.

beta: f64

Trend smoothing factor β.

level: f64

Last smoothed level.

trend: f64

Last smoothed trend.

phi: f64

Damping factor ϕ (1.0 = undamped).

residual_std: f64

Standard deviation of in-sample residuals.

HoltWinters

Holt-Winters (triple exponential smoothing) parameters.

Fields

alpha: f64

Level smoothing factor α.

beta: f64

Trend smoothing factor β.

gamma: f64

Seasonal smoothing factor γ.

period: usize

Seasonal period length (number of observations per cycle).

level: f64

Last smoothed level.

trend: f64

Last smoothed trend.

seasonal: Vec<f64>

Seasonal indices (one per period position).

multiplicative: bool

Whether multiplicative (true) or additive (false) seasonality is used.

residual_std: f64

Standard deviation of in-sample residuals.

seasonal_phase: usize

Phase offset into the seasonal vector for the first prediction step. Set to n_train % period during fit so predictions continue the cycle.

Arima

ARIMA(p,d,q) parameters.

Fields

p: usize

Autoregressive order.

d: usize

Differencing order.

q: usize

Moving average order.

ar_coeffs: Vec<f64>

Autoregressive coefficients φ₁…φₚ.

ma_coeffs: Vec<f64>

Moving average coefficients θ₁…θ_q.

constant: f64

Mean of the differenced series, added back to every forecast.

Zero when the model carries no constant — see ArimaModel for when one is included.

residual_std: f64

Standard deviation of in-sample residuals.

effective_ar_order: Option<usize>

Effective AR order after Burg estimation. May be less than p if the algorithm terminated early due to near-constant input data.

Sarima

Seasonal ARIMA (p,d,q)(P,D,Q)m parameters.

Fields

p: usize

Non-seasonal AR order.

d: usize

Non-seasonal differencing order.

q: usize

Non-seasonal MA order.

sp: usize

Seasonal AR order P.

sd: usize

Seasonal differencing order D.

sq: usize

Seasonal MA order Q.

m: usize

Seasonal period m.

ar_coeffs: Vec<f64>

Non-seasonal AR coefficients.

ma_coeffs: Vec<f64>

Non-seasonal MA coefficients.

sar_coeffs: Vec<f64>

Seasonal AR coefficients Φ₁…Φ_P.

sma_coeffs: Vec<f64>

Seasonal MA coefficients Θ₁…Θ_Q.

constant: f64

Mean of the fully differenced series, added back to every forecast.

residual_std: f64

Standard deviation of in-sample residuals.

LinearRegression

Linear regression parameters.

Fields

slope: f64

Regression slope (change per unit index).

intercept: f64

Regression intercept.

r_squared: f64

Coefficient of determination R².

residual_std: f64

Standard deviation of in-sample residuals.

start_ts: i64

Start timestamp of training data (nanoseconds).

interval_ns: i64

Observation interval (nanoseconds).

Custom

Opaque parameters for a custom plugin model.

Fields

name: String

Plugin model name.

data: Vec<u8>

Opaque serialized parameter data (interpreted by the plugin).

Trait Implementations

impl Clone for ModelParams

fn clone(&self) -> ModelParams

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for ModelParams

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for ModelParams

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for ModelParams

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.