# smooth

> R and Python implementations of state-space forecasting models in the
> Single Source of Error (SSOE) framework. Primary function: ADAM
> (Augmented Dynamic Adaptive Model), unifying ETS, ARIMA, and regression.
> R package on CRAN, Python port on PyPI. Source:
> https://github.com/config-i1/smooth. ADAM book:
> https://openforecast.org/adam/.

Status legend used below: `stable` = implemented and in regular use;
`R-only` = R has it, Python port pending (see Roadmap); `partial` = exists
in Python but missing features compared to R.

## Navigation

- [Home](https://github.com/config-i1/smooth/wiki/Home): wiki landing page with the function index, parameter index, and R↔Python parity matrix.
- [Glossary](https://github.com/config-i1/smooth/wiki/Glossary): disambiguation of overloaded terms (`ar` vs `ar_order` vs `AR` vs `"ar"`, ETS letter codes, state-component names).
- [Roadmap](https://github.com/config-i1/smooth/wiki/Roadmap): functionality that exists in R smooth but is not yet in the Python port.
- [R-Python-differences](https://github.com/config-i1/smooth/wiki/R-Python-differences): numerical-parity status between the R and Python implementations.
- [Installation](https://github.com/config-i1/smooth/wiki/Installation): install instructions for R (CRAN) and Python (PyPI), plus build-from-source notes.
- [Resources](https://github.com/config-i1/smooth/wiki/Resources): academic references and DOIs.

## Forecasting functions

- [ADAM](https://github.com/config-i1/smooth/wiki/ADAM) — `adam()` (R) / `ADAM` (Python). Stable. Unified ETS + ARIMA + regression in SSOE form.
- [AutoADAM](https://github.com/config-i1/smooth/wiki/AutoADAM) — `auto.adam()` / `AutoADAM`. Stable. Automatic distribution and ARIMA-order selection.
- [ES](https://github.com/config-i1/smooth/wiki/ES) — `es()` / `ES`. Stable. Exponential Smoothing wrapper for ADAM (pure ETS, no ARIMA).
- [CES](https://github.com/config-i1/smooth/wiki/CES) — `ces()`, `auto.ces()` / `CES`, `AutoCES`. Stable. Complex Exponential Smoothing.
- [MSARIMA](https://github.com/config-i1/smooth/wiki/MSARIMA) — `msarima()`, `auto.msarima()` / `MSARIMA`, `AutoMSARIMA`. Stable. Multiple Seasonal ARIMA.
- [SSARIMA](https://github.com/config-i1/smooth/wiki/SSARIMA) — `ssarima()`. R-only. Pure SSARIMA; use MSARIMA in Python.
- [GUM](https://github.com/config-i1/smooth/wiki/GUM) — `gum()`. R-only. Generalised Univariate Model with user-supplied F / w / g.
- [SMA](https://github.com/config-i1/smooth/wiki/SMA) — `sma()` / `SMA`. Stable. Simple Moving Average in state-space form.
- [OM](https://github.com/config-i1/smooth/wiki/OM) — `om()`, `omg()`, `auto.om()`, `oes()` / `OM`, `OMG`, `AutoOM`. Stable. Occurrence model for intermittent demand.
- [OES](https://github.com/config-i1/smooth/wiki/OES) — `oes()`. Stable. Older oETS entry point; the OM page is the current preferred reference.
- [msdecompose](https://github.com/config-i1/smooth/wiki/msdecompose) — `msdecompose()` (R and Python). Stable. Multiple seasonal decomposition used for initialisation.

## Parameter pages (shared across functions)

- [Model-Specification](https://github.com/config-i1/smooth/wiki/Model-Specification) — the `model` parameter, ETS letter codes, automatic-selection codes (`ZZZ`/`ZXZ`/`XXX`/`YYY`/`FFF`/`PPP`/`SSS`/`CCC`), model-type extractors.
- [Orders-and-Lags](https://github.com/config-i1/smooth/wiki/Orders-and-Lags) — `orders`, `lags`, and the Python scalar trio `ar_order` / `i_order` / `ma_order` for ARIMA components and seasonal periods.
- [Loss-Functions](https://github.com/config-i1/smooth/wiki/Loss-Functions) — the `loss` parameter: likelihood, MSE, MAE, HAM, GTMSE, LASSO, RIDGE, custom callables, and multi-step variants.
- [Bounds](https://github.com/config-i1/smooth/wiki/Bounds) — the `bounds` parameter: `"usual"`, `"admissible"`, `"none"`, and stability conditions on smoothing parameters.
- [Initialisation](https://github.com/config-i1/smooth/wiki/Initialisation) — the `initial` parameter: backcasting, optimal, two-stage, complete, or user-provided initial states.
- [Persistence](https://github.com/config-i1/smooth/wiki/Persistence) — smoothing parameters alpha / beta / gamma / delta and the damping `phi`.
- [Explanatory-Variables](https://github.com/config-i1/smooth/wiki/Explanatory-Variables) — external regressors via `formula`/`xreg` (R) and `X` (Python); the `regressors` mode.
- [Model-Estimation](https://github.com/config-i1/smooth/wiki/Model-Estimation) — optimiser, information criterion, Fisher Information / Hessian, advanced estimation options.

## Method / extractor pages (called on fitted models)

- [Fitted-Values-and-Forecasts](https://github.com/config-i1/smooth/wiki/Fitted-Values-and-Forecasts) — `fitted()`, `actuals()`, `forecast()` / `predict()`.
- [Coefficients-and-Parameters](https://github.com/config-i1/smooth/wiki/Coefficients-and-Parameters) — `coef()`, `confint()`, `vcov()`, `coefbootstrap()`.
- [Residuals-and-Errors](https://github.com/config-i1/smooth/wiki/Residuals-and-Errors) — `residuals()`, `rstandard()`, `rstudent()`, `rmultistep()`, `multicov()`, `outlierdummy()`.
- [Likelihood-and-Information-Criteria](https://github.com/config-i1/smooth/wiki/Likelihood-and-Information-Criteria) — `logLik()`, `AIC` / `BIC` / `AICc` / `BICc`, `accuracy()`, `pls()`.
- [Model-Information](https://github.com/config-i1/smooth/wiki/Model-Information) — `nobs()`, `nparam()`, `sigma()`, `modelType()`, `modelName()`, `orders()`, `lags()`.
- [Visualisation-and-Output](https://github.com/config-i1/smooth/wiki/Visualisation-and-Output) — `print()`, `summary()`, `plot()`, `xtable()` (R-only).
- [Refitting-and-Reforecasting](https://github.com/config-i1/smooth/wiki/Refitting-and-Reforecasting) — `reapply()`, `reforecast()`.
- [Scale-Model](https://github.com/config-i1/smooth/wiki/Scale-Model) — `sm()`. R-only. Models the scale (variance) of the error term dynamically.
- [Simulation-Functions](https://github.com/config-i1/smooth/wiki/Simulation-Functions) — `sim.*` (R) / `sim_*` (Python) family plus the `.simulate()` method on fitted models.