Learn R Programming

'mixAR' is an R package for time series modelling with mixture autoregressive and related models.

Installing mixAR

Install the latest stable version from CRAN:

install.packages("mixAR")

Alternatively, install the development version of mixAR from Github:

remotes::install_github("GeoBosh/mixAR")

Overview

Package mixAR provides functions for modelling with mixture autoregressive (MAR/mixAR) models.

See, for example, the overview of mixAR and model fitting functions such as fit_mixAR and bayes_mixAR.

Copy Link

Version

Install

install.packages('mixAR')

Monthly Downloads

267

Version

0.22.9

License

GPL (>= 2)

Homepage

https://geobosh.github.io/mixAR/ (doc)

Maintainer

Georgi Boshnakov

Last Published

December 16th, 2025

Functions in mixAR (0.22.9)

MixARGaussian-class

mixAR models with Gaussian noise components

Bayesian sampling of mixture autoregressive models

Class "MixARgen"

MixVARGaussian-class

MixVAR models with multivariate Gaussian noise components

Adjust the length of the second argument to be the same as that of the first one

Class "MixVAR" --- mixture vector autoregressive models

Class "MixComp" --- manipulation of MixAR time series

Class "MixAR" --- mixture autoregressive models

Closing prices of four stocks

Optimise scale parameters in MixARgen models

RJMCMC move for AR order selection of mixture autoregressive models

Functions for the standard normal distribution

Update the scale parameters of MixAR models

Create estimation templates from MixAR model objects

Log-likelihood of MixAR models

companion_matrix

Create a companion matrix from a vector

MixAR models for examples and testing

Gaussian EM-step with random initialisation

Choose the autoregressive order of MixAR components

Utility function for MixAR

fit_mixAR-methods

Fit mixture autoregressive models

Internal mixAR Functions

get_edist-methods

Methods for function get_edist in package mixAR

Calculate component specific error terms under MixAR model

Create MixAR objects

initialize-methods

Methods for function initialize in package mixAR

make_fcond_lik-methods

Create a function for computation of conditional likelihood

Generator functions for noise distributions

Generalised inner product and methods for class "MixComp"

Fit mixture autoregressive models with seasonal AR parameters

Check if a MixAR model is stable

Internal functions for estimation of MixAR models with Gaussian components

A posteriori relabelling of a Markov chain

mixARemFixedPoint

EM estimation for mixture autoregressive models

Vector of parameters of a MixAR model

fit_mixARreg-methods

Fit time series regression models with mixture autoregressive residuals

Fit mixture vector autoregressive models

Relabel the components of a MixAR model

Extract the last n elements of a vector

mix_location-methods

Conditional mean of MixAR models

Simulate white noise series from a list of functions and vector of regimes

Compute probabilities for the observations to belong to each of the components

Diagnostic checks for mixture autoregressive models

mix_central_moment-methods

Methods for mix_central_moment

fit_mixVAR-methods

Fit mixture vector autoregressive models

lik_params_bounds

Give natural limits for parameters of a MixAR model.

Calculate marginal loglikelihood at high density points of a MAR model.

Filter time series with MixAR filters

M-step for models from class MixARgen

Simulate from MixAR models

multiStep_dist-methods

Multi-step predictions for MixAR models

require("mixAR") pd <- packageDescription("mixAR") lb <- library(help="mixAR", character.only=TRUE) lbinfo <- if(!is.null(lb$info[[2]])) lb$info[[2]] else "No_matches" anypat <- function(string, pat, ...){ any(sapply(pat, function(x) grepl(x, string, ...))) } lbsubset <- function(lbinfo, pat){ if(length(lbinfo) == 0) return("No entries") ind.start <- which(sapply(lbinfo, function(x) anypat(x, pat))) if(length(ind.start) == 0) return("No entries") ind.ws <- which(grepl("^[[:space:]]+", lbinfo)) res <- ind.start for(ind in ind.start){ while(any((ind+1) - ind.ws == 0)){ ind <- ind + 1 res <- c(res, ind) } } lbinfo[ sort(res) ] } lbpaste <- function(pat) paste("\\preformatted{", {wrk <- lbsubset(lbinfo,pat); wrk <- gsub("(^[^\\]?%)|(([^\\])%)", paste0("\\3", "\\\\", "%"), wrk); paste(if(length(wrk) > 0) wrk else "No entries", collapse="\n")}, "}", sep="") pd$Title

mix_moment-methods

Methods for mix_moment

Internal mixAR functions

noise_dist-methods

Methods for function noise_dist in package mixAR

Infix operator to apply functions to matrix-like objects

mixARExperiment

Simulation experiments with MixAR models

mix_ncomp-methods

Number of rows or columns of a MixComp object

raggedCoefS-class

Class "raggedCoefS" --- ragged list

randomArCoefficients

Random initial values for MixAR estimation

BIC based model selection for MixAR models

noise_rand-methods

Methods for function noise_rand in package mixAR

mix_variance-methods

Methods for mix_variance

Function and methods to compute component residuals for MixAR models

All permutations of the columns of a matrix

Compute moments and absolute moments of standardised-t and normal distributions

Estimate probabilities of a MixAR model from tau.

raggedCoefV-class

Class "raggedCoefV" --- ragged list

Exact predictive parameters for multi-step MixAR prediction

Compute component predictions for MixAR models

Small utilities for ragged objects

mixAR_cond_probs

The E-step of the EM algorithm for MixAR models

Filter a time series with options to shift and scale

Dealing with label switching in MixAR experiments

Support for EM estimation of MixAR models, internal function.

Set or extract the parameters of MixAR objects

Compute standard errors of estimates of MixAR models

Show differences between two models

Convert a ragged list into a matrix of characters

Conditional moments of MixAR models

noise_moment-methods

Compute moments of the noise components

Conditional quantile functions of MixAR models

noise_params-methods

Methods for function noise_params in package mixAR

Simulate from multivariate MixAR models

mix_pdf-methods

Conditional pdf's and cdf's of MixAR models

Put core MixAR coefficients into a canonical form, internal function

row_lengths-methods

Methods for function row_lengths in package mixAR

Sampling functions for Bayesian analysis of mixture autoregressive models

Perform simulation experiments

Utility function for mixAR

tomarparambyComp

Translations of my old MixAR Mathematica functions

raggedCoef-class

Class "raggedCoef" --- ragged list objects

A test for 'unswitch'