openforecast-org

Explain and interpret smooth (ADAM) state-space forecasting outputs in plain language, and pick the right model function — ADAM/AutoADAM, ES, CES/AutoCES, MSARIMA/AutoMSARIMA, SMA, the occurrence models OM/OMG/AutoOM for intermittent demand, msdecompose, and the sim_* simulators. Covers ETS model notation (the three-letter code and Z/X/Y/C/F selection placeholders), persistence/smoothing parameters (alpha, beta, gamma, phi) and their constraints, ARIMA orders, error distributions, information-criteria model selection, point forecasts and prediction intervals, component/state decomposition, holdout accuracy, and explanatory variables / external regressors (ETSX / ARIMAX / oETSX) — the `formula`/`xreg` (R) and `X` (Python) arguments, the `regressors` mode (`use`/`select`/`adapt`/`integrate`), and the Python intercept-drop and `adapt`-bounds caveats. Use when the user asks what a fitted model means, how to read a summary/forecast/plot, why a model or distribution was selected, how to add regressors, or which fun

Explain and interpret greybox outputs in plain language — ALM model summaries (coefficients, sigma, information criteria), confidence/prediction intervals, forecast accuracy measures, stepwise/CALM selection, RMCB/Nemenyi method comparisons, stick() STI decomposition, distributions, and association measures. Use when the user asks what a result means, how to read a summary or plot, which metric/distribution to pick, or why a model behaves a certain way — in either the R package or the Python port.