ets

Source: src/ets.rs

Exponential smoothing forecasters: SES, Holt, Holt-Winters.

Pure numerical core — accepts a slice of f64 values and parameters, returns a Vec of h forecast values. The Python wrapper handles grouping, time columns, and DataFrame construction.

Python API

ets_ses

ets_ses(values, alpha, h)

Rust-accelerated Simple Exponential Smoothing.

Takes values for a single group and returns h forecast values.

ets_holt

ets_holt(values, alpha, beta, h)

Rust-accelerated Holt's linear trend method.

ets_holt_winters

ets_holt_winters(values, alpha, beta, gamma, season_length, additive, h)

Rust-accelerated Holt-Winters seasonal method.

Internal Functions

These are internal Rust functions not directly callable from Python.

ses_core

Simple Exponential Smoothing.

Smooths with level parameter alpha and projects a flat forecast of length h.

holt_core

Holt's linear trend method.

Smooths level and trend, then extrapolates linearly for h steps.

holt_winters_core

Holt-Winters seasonal method (additive or multiplicative).

Smooths level, trend, and seasonal components, then forecasts h steps.