model_combination_ops: Combine models jointly via sum, mean, or weighted mean

In fabletools: Core Tools for Packages in the 'fable' Framework

Combine models jointly via sum, mean, or weighted mean

Description

These methods allow multiple fitted models to be combined into a single joint combination model. Unlike the pairwise arithmetic approach (e.g. (m1 + m2 + m3)/3), these solve the N-way convolution of all component distributions simultaneously using the full residual covariance matrix, so the result is invariant to the order in which models are listed.

Usage

## S3 method for class 'mdl_ts'
Summary(..., na.rm = FALSE)

## S3 method for class 'mdl_ts'
mean(x, ...)

## S3 method for class 'mdl_ts'
weighted.mean(x, w, ...)

## S3 method for class 'mdl_lst'
Summary(..., na.rm = FALSE)

## S3 method for class 'mdl_lst'
mean(x, ...)

## S3 method for class 'mdl_lst'
weighted.mean(x, w, ...)

Arguments

...	Additional models to include in the combination.
na.rm	Ignored; present for compatibility with the Summary generic.
x	The first model (or a list/mdl_lst of models).
w	A numeric vector of weights, one per model (will be normalised to sum to 1).

Value

A fitted combination model (mdl_ts) of class model_combination.

See Also

combination_ensemble(), combination_weighted()

fabletools documentation built on June 28, 2026, 5:06 p.m.