STAR_outliers: a python package that separates univariate outliers from non-normal distributions

Table 2 A list of algorithms compared to STAR_outliers. Algorithms are detailed in the figure generation repository

Algorithm	Algorithm type	Description
IF	IF	IF out of the box model
IF-calibrated	IF	IF calibrated to remove as many outliers as STAR_outliers
STAR	STAR_outliers	STAR_outliers
[3] (p = 90)	[3]	[3] using percentiles 90 and 10 to estimate Tukey-gh parameters
[3] (p = 91)	[3]	[3] using percentiles 91 and 9 to estimate Tukey-gh parameters
[3] (p = 92)	[3]	[3] using percentiles 92 and 8 to estimate Tukey-gh parameters
[3] (p = 93)	[3]	[3] using percentiles 93 and 7 to estimate Tukey-gh parameters
[3] (p = 94)	[3]	[3] using percentiles 94 and 6 to estimate Tukey-gh parameters
[3] (p = 95)	[3]	[3] using percentiles 95 and 5 to estimate Tukey-gh parameters
[3] (p = 96)	[3]	[3] using percentiles 96 and 4 to estimate Tukey-gh parameters
[3] (p = 97)	[3]	[3] using percentiles 97 and 3 to estimate Tukey-gh parameters
[3] (p = 98)	[3]	[3] using percentiles 98 and 2 to estimate Tukey-gh parameters
[3] (p = 99)	[3]	[3] using percentiles 99 and 1 to estimate Tukey-gh parameters
T2	[9]	T2 with 2 iterations
T2_yj	[9]	T2 with 2 iterations for yj transformed data
T3	[9]	T2 with 3 iterations
T3_yj	[9]	T2 with 3 iterations for yj transformed data
T4	[9]	T2 with 4 iterations
T4_yj	[9]	T2 with 4 iterations for yj transformed data
3SD	normal	standard 3SD cutoff
3SD_yj	normal	3SD cutoff for yj transformed data
IQR	normal	standard IQR cutoff
IQR_yj	normal	standard IQR cutoff for yj transformed data
MAD	normal	standard MAD cutoff
MAD_yj	normal	standard MAD cutoff for yj transformed data

ISSN: 1756-0381