omg: General occurrence model

In smooth: Forecasting Using State Space Models

General occurrence model

Description

Fits two parallel ETS occurrence models (A: odds-ratio, B: inverse-odds-ratio) jointly using a shared Bernoulli log-likelihood. The combined probability at each time point is p_t = p_{At} / (p_{At} + p_{Bt}).

Usage

omg(data, modelA = "MNN", modelB = modelA, ordersA = list(ar = c(0), i =
  c(0), ma = c(0), select = FALSE), ordersB = ordersA, constantA = FALSE,
  constantB = constantA, formulaA = NULL, formulaB = formulaA,
  regressorsA = c("use", "select", "adapt"), regressorsB = regressorsA,
  persistenceA = NULL, persistenceB = persistenceA, phiA = NULL,
  phiB = phiA, armaA = NULL, armaB = armaA, etsA = c("conventional",
  "adam"), etsB = etsA, lags = c(frequency(data)), h = 0,
  holdout = FALSE, initial = c("backcasting", "optimal", "two-stage",
  "complete"), loss = c("likelihood", "MSE", "MAE", "HAM", "LASSO", "RIDGE"),
  ic = c("AICc", "AIC", "BIC", "BICc"), bounds = c("usual", "admissible",
  "none"), model = NULL, silent = TRUE, ...)

Arguments

data	Binary time series (0/1), vector or data frame.
modelA	ETS model string for model A (default "MNN").
modelB	ETS model string for model B (default "MNN").
ordersA	ARIMA orders list for model A.
ordersB	ARIMA orders list for model B.
constantA	Logical, include constant in model A.
constantB	Logical, include constant in model B.
formulaA	Formula for exogenous variables in model A.
formulaB	Formula for exogenous variables in model B.
regressorsA	How to handle regressors in model A.
regressorsB	How to handle regressors in model B.
persistenceA	Persistence vector for model A.
persistenceB	Persistence vector for model B.
phiA	Damping parameter for model A.
phiB	Damping parameter for model B.
armaA	ARMA parameters for model A.
armaB	ARMA parameters for model B.
etsA	ETS variant for model A ("conventional" or "adam").
etsB	ETS variant for model B.
lags	Seasonal lags (shared).
h	Forecast horizon.
holdout	If TRUE, hold out the last h observations.
initial	Initialisation method (shared).
loss	Loss function (shared).
ic	Information criterion (shared).
bounds	Parameter bounds type (shared).
model	An already-fitted omg object. When supplied, the per-side parameters are lifted from model$modelA and model$modelB and no estimation is performed; passing FI=TRUE alongside computes the observed Fisher information over the joint parameter vector (the path used by vcov.omg).
silent	If TRUE, suppress output.
...	Additional arguments passed to the optimiser.

Value

An object of class c("omg","om","smooth").

See Also

om, forecast.omg

Examples

set.seed(41)
y <- rpois(100, 0.5)
m <- omg(y)
forecast(m, h=10)

smooth documentation built on June 21, 2026, 9:06 a.m.