tsensemble: Ensembling of Simulations

In tsaux: Time Series Forecasting Auxiliary Functions

Ensembling of Simulations

Description

Ensembling of Simulations

Usage

## S3 method for class 'tssim.mixture'
tsensemble(object, weights = NULL, difference = TRUE, ...)

Arguments

object	an object of class tssim.mixture.
weights	the weighting (or probability) matrix for aggregating the simulations (see details).
difference	whether to take the rates of changes first before aggregating and reconverting to levels.
...	additional parameters.

Details

When mixing dynamics for the same series, and when series are not stationary, differences should be used. In that case the rate of change transformation is applied to each simulated series and then weighted by the weights matrix. Since the weights matrix will have one more row than is required (the first row), this can be used to choose how the initial level is generated. For instance, if we want to use the level of the first simulated series, then the first row would have a 1 on the first column and zeros in the rest. For aggregating series, difference should be set to FALSE since we are looking at summation of data (under the assumption of flow variables). In this case, the p matrix is usually static by column (i.e. the same weights).

Value

A vector of the simulated series.

tsaux documentation built on April 4, 2025, 3:08 a.m.