ForecastGibbs: Update the design matrix with a forecasted data point.

In gamalamboy/stresstest: Implement the ETI Stress Testing Simulation

Description

Given a model structure and a OneStep object to use as new data value, create a new design matrix Z_{t} from Z_{t-1}.

Usage

ForecastGibbs(dats_withexo, period, ztprop1, gibbs_mean, gibbs_var)

Arguments

dats_withexo	the dataset used to create the forecast. It consists of an enlarged zoo object with NA's in the forecast rows (except for exogenous variables). The dataset must be generated by DataInput().
period	the number of years to forecast. As of yet, the function does not support backcasting.
ztprop1	the model structure (a Ztprops object) to be used in determining the proper positions of each value.
gibbs_mean	the Mean component of the Gibbs simulation.
gibbs_var	the Var component of the Gibbs simulation.

Details

this function operates in three (3) parts:

1. shift old values of the variables by 1 using the lag operator.

2. reinsert shifted values in the proper positions of Z_{t}.

3. replace lag 1 values with values from the OneStep object.

This produces a design matrix Z_{t} for the forecasted year.

Value

an array with dimensions (<number_of_forecasted_years>, <number_of_equations>, <number_of_iterations_in_Gibbs>).

gamalamboy/stresstest documentation built on May 17, 2019, 1:33 p.m.