run_model: Get nowcast from the bayesian DFM

In h4sci/packagr: Estimation of a Bayesian Dynamic Factor Model

Description

function extends the data matrix by the forecasting period, draws extended factor conditional on posterior parameters and makes forecast.

Usage

run_model(yt, k, q, m, n, Tt, Ttq, const, target, inventory, Xmat, flows)

Arguments

yt	A matrix of demeaned and standardized time series data.
k	Number of states (number of factors)
q	Lag length for state equation.
m	number of states (number of factors) x lag length for state equation (nxq)
n	Number of variables
Tt	Number of high-frequency periods.
Ttq	Number of high-frequency periods minus lag length for state equation
const	A scalar, where const = 1 for model with intercept, const = 0 for model without intercept.
target	Defines the target variable.
inventory	Inventory of the corresponding data.
Xmat	A data set containing all the time series.
flows	A list of time series variable (flow variables).

Value

A list with the mean and quantiles of the parameters, the forecast and the number of burns, draws and saves.

Examples

yt <- as.matrix(t(Xmat))
flows = data$flows
inventory <- create_inventory(flows = data$flows, stocks = data$stocks)
k <- 2
n <- dim(yt)[1]
q <- 1
const <- 0
m <- k*q
Tt <- dim(yt)[2]
Ttq <- Tt-q
target <- c("UKGDPM.YQ")
out <- run_model(yt,k,q,m,n,Tt,Ttq,const,target,inventory,Xmat,flows)

h4sci/packagr documentation built on Jan. 7, 2021, 10:40 p.m.