SHASH_out: Robust outlier detection based on SHASH distribution

In fMRIscrub: Scrubbing and Other Data Cleaning Routines for fMRI

Robust outlier detection based on SHASH distribution

Description

A robust outlier detection based on modeling the data as coming from a SHASH distribution.

Usage

SHASH_out(x, maxit = 20, out_lim = 3, weight_init = NULL)

Arguments

x	The numeric vector in which to detect outliers.
maxit	The maximum number of iterations. Default: 10.
out_lim	SD threshold for outlier flagging. Default: 4.
weight_init	Initial weights. Default: NULL (no pre-determined outliers).

Value

A "SHASH_out" object, i.e. a list with components

out_idx

Indices of the detected outliers.

x_norm

The normalized data.

SHASH_coef

Coefficients for the SHASH-to-normal transformation.

indx_iters

TRUE for the detected outliers for each itertation.

last_iter

Last iteration number.

converged

Logical indicating whether the convergence criteria was satisfied or not.

Examples

x <- rnorm(100) + (seq(100)/200)
x[77] <- 13
SHASH_out(x)

fMRIscrub documentation built on Oct. 25, 2023, 9:07 a.m.