dfmMS: Dynamic factor model with Markov-switching states

Description Usage Arguments Value

Description

Dynamic factor model with Markov-switching states

Usage

1
dfmMS(X, J = 1, s = 2, p = 1, x0, P0)

Arguments

X

Data matrix (Txn)

J

Number of factors. Currently, only 1 is supported

s

Number of states. Only 2 are supported for now

p

Lag number for VAR model of factors

x0

Initial value for state

P0

Initial value for state covariance

Value

Return an estimator. Currently, it runs via a likelihood maximization an so is rather slow. Do not call it with large number of input variables or missing data. Some box constraints are currently to speed up the process, these are based on usual rationale for economic data.


guilbran/dynfactoR documentation built on May 8, 2019, 1:35 a.m.