Table 3 A list of distributions with simulated outliers analyzed by the algorithms in Table 2. Nonstandard distributions include uniform distributions, triangular distributions, multimodal normal distributions, and different mixtures thereof. They are defined in step1 of the figure generation repository
From: STAR_outliers: a python package that separates univariate outliers from non-normal distributions
distribution Type | Description |
|---|---|
lognormal | lognormal distribution: 10 spread parameter values |
exponential | exponential random variables drawn and then raised to a power: 10 power values |
power | power distribution: 10 power values |
poisson | poisson distribution: 10 parameter values |
negative binomial | negative binomial distribution: 10 success probabilities |
Tukey-g | tukey-gh distribution: 10 g values, h = 0 |
Tukey-h | tukey-gh distribution: 10 h values, g = 0 |
Tukey-gh | tukey-gh distribution: 10 values, g = h |
beta | beta distribution: 10 beta values, alpha = 2 |
non-standard | various non-smooth and/or multimodal distributions: 10 different shapes |