arma_est_bip_mm: arma_est_bip_mm

In Mufabo/Rrobustsp: Robust Signal Processing

arma_est_bip_mm

Description

The function arma_est_bip_mm(x,p,q) comuptes BIP MM-estimates of the ARMA model parameters. It also computes an outlier cleaned signal using BIP-ARMA(p,q) predictions

Usage

arma_est_bip_mm(x, p, q, tolX = 5e-07)

Arguments

x:	data (observations/measurements/signal)
p:	autoregressive order
q:	moving-average order

Value

result: named list with following fields

ar_coeffs	numeric vector of length p. BIP-AR(p) MM estimates
ma_coeffs	numeric vector of length q. BIP-AR(q) MM estimates
inno_scale	numeric, BIP s-estimate of the innovations scale
ar_coeffs_init	numeric vector of length p. Robust starting point for estimation
ma_coeffs_init	numeric vector of length q. Robust starting point for estimation

Note

file is in dependentData_armaEstBipMM.R

References

"Robust Statistics for Signal Processing" Zoubir, A.M. and Koivunen, V. and Ollila, E. and Muma, M. Cambridge University Press, 2018.

"Bounded Influence Propagation τ-Estimation: A New Robust Method for ARMA Model Estimation." Muma, M. and Zoubir, A.M. IEEE Transactions on Signal Processing, 65(7), 1712-1727, 2017.

Examples

library(signal)
library(zeallot)
library(pracma)

N <- 500
a <- rnorm(N)
p <- 1
q <- 0
x <- signal::filter(1, c(1, -0.8), a)

arma_est_bip_mm(x, p, q)

Mufabo/Rrobustsp documentation built on June 11, 2022, 10:41 p.m.