RMSDp {RMSDp}

Modified Stahel-Donoho Estimators (parallel version)

Description

This function is for multivariate outlier detection. version 0.0.1 2013/06/15 Related paper: DOI: 10.1109/CLOUDCOM-ASIA.2013.86 version 0.0.2 2021/11/15 Outlier detection step added version 0.0.3 2022/08/12 Bug fixed about Random seed setting

Usage

RMSDp(inp, cores = 0, nb = 0, sd = 0, pt = 0.999, dv = 10000)

Arguments

inp

input data (a numeric matrix)

cores

number of cores used for this function

nb

number of basis

sd

seed (for reproducibility)

pt

threshold for outlier detection (probability)

dv

maximum number of elements processed together on the same core

Value

a list of the following information

• u final mean vector

• V final covariance matrix

• wt final weights

• mah squared squared Mahalanobis distances

• cf threshold to detect outlier (percentile point)

• ot outlier flag (1:normal observation, 2:outlier)