KimFilter: Implementation of Kim (1994) filter, an extension to Kalman...

In SebKrantz/dynfacto_R: Dynamic Factor Model Estimation for Nowcasting

Description

Implementation of Kim (1994) filter, an extension to Kalman filter for dynamic linear models with Markov-switching. Documentation is incomplete, rudimentary and needs to be rechecked!

Kim filtering

Usage

KimFilter(y, R, Q, F1, A1, x0, P0, p)

KimFilter(y, R, Q, F1, A1, x0, P0, p)

Arguments

y	Data matrix (Txn)
R	Error covariance for measurement equation
Q	Error covariance for state equation
F1	Array of observation matrices, one matrix per state
A1	Array of transition matrices, one matrix per state
x0	Initial condition for state vector
P0	Initial condition for state variance
p	Transition probability matrix
F	System matrix for measurement equation
A	Transition matrix for state equation

Value

Filtered states and covariances with associated probability matrices.

This function returns the same output as KimFilterCpp. However, it converts certain elements of the Kim Filter to standard multi-dimensional R arrays and so its outputs can be used as inputs for KimSmoother.

See Also

KimFilterCpp

SebKrantz/dynfacto_R documentation built on Dec. 31, 2020, 4:30 p.m.