Fincept

ARIMA / SARIMAX forecast

stats_arima

Read-onlyIdempotent

ARIMA / seasonal SARIMA with optional exogenous regressors (ARIMAX/SARIMAX), fitted by maximum likelihood, forecasting steps periods ahead with confidence intervals (series + line chart with band). Returns the coefficient table, AIC/BIC/HQIC/log-likelihood, Ljung-Box residual checks and optional fitted values. Use it as the default forecaster for one series; stats_arma_order_select picks p and q, stats_ets and stats_theta are smoothing alternatives.

Input Schema

TableJSON Schema

Name	Required	Description	Default
`y`	Yes	Numbers in order (null for a missing value), or a $fincept reference.
`exog`	No	Exogenous regressors, one row per observation of y, columns by name (e.g. rates, prices, dummies).
`alpha`	No	1 - confidence of intervals (0.05 = 95% intervals).
`dates`	No	Optional ISO dates, one per observation; gives forecasts and periods real dates.
`order`	No	Non-seasonal order [p, d, q]: AR lags, differences, MA lags.
`steps`	No	Periods to forecast after the last observation (0 = none).
`trend`	No	Deterministic terms: n none, c constant, t linear trend, ct both. Default c without differencing, n with. With estimator sarimax and d=1, c is a drift; with the other estimators use t for a drift.
`estimator`	No	sarimax: exact Gaussian likelihood in state space (handles gaps, seasonality, regressors). arima_mle: the same likelihood as a regression with ARIMA errors. innovations_mle, hannan_rissanen: alternative ARMA estimators. burg, yule_walker: pure AR only (q=0). innovations: pure MA only (p=0).	sarimax
`future_exog`	No	Regressor values over the forecast horizon: one row per step, the same columns as exog. Required when exog is given and steps > 0.
`include_fitted`	No	Also return in-sample fitted values and residuals for every observation.
`seasonal_order`	No	Seasonal order [P, D, Q, s]: seasonal AR, seasonal differences, seasonal MA, and the period s (12 monthly, 4 quarterly, 5 or 7 daily).