marx: The MARX function
In MARX: Simulation, Estimation, Model Selection and Forecasting for MARX Models

Description Usage Arguments Details Value Author(s) Examples

This interface-based function allows you to perform model selection for MARX models based on information criteria.

1	marx(y, x, p_max, sig_level, p_C, p_NC)

`y`	Data vector of time series observations.
`x`	Matrix of data (every column represents one time series). Specify NULL or "not" if not wanted.
`p_max`	Maximum number of autoregressive parameters (leads + lags) to be included.
`sig_level`	Significance level for the construction of inference.
`p_C`	Number of lags (if not specified by the user a model selection procedure is used to determine the number of lags).
`p_NC`	Number of leads (if not specified by the user a model selection procedure is used to determine the number of leads).

Mixed causal-noncausal autoregressions with exogenous regressors.

The function returns the values of the information criteria for the pseudo-causal models. The user is asked to choose a value for "p". Extensive output for the MARX(r,s,q) model (with p = r + s) which maximizes the log-likelihood is reported.

Sean Telg

data <- sim.marx(c('t',1,1), c('t',1,1),100,0.5,0.4,0.3)
p_max <- 8
sig_level <- 0.05
marx(data$y, data$x, p_max, sig_level,1,1) ## p_C and p_NC chosen to be 1: MARX(1,1,1) output.
marx(data$y, NULL, p_max,sig_level,1,1) ## MAR(1,1), no exogenous variable specified.