generalized_ESD: Detect outliers using the generalized Extreme Studentized...
In skinnider/modern: Model-free outlier detection for robust network inference

Description Usage Arguments Details Value

View source: R/generalized_ESD.R

Identify outliers within a distribution of numeric values using Rosner's generalized Extreme Studentized Deviate (ESD) test. Unlike the Grubbs or Tietjen-Moore tests, the generalized ESD test does not require the number of outliers to be prespecified.

1 2	generalized_ESD(x, max_outliers = 5, alpha = 0.05, return_statistics = T)

`x`	distribution to find outliers in
`max_outliers`	the prespecified maximum number of outliers; defaults to `5`
`alpha`	type I error associated with the hypothesis test; defaults to `0.05`
`return_statistics`	optionally return the test statistics associated with the top `max_outliers` outliers

Full details are provided in:

Rosner, Bernard (May 1983), Percentage Points for a Generalized ESD Many-Outlier Procedure, Technometrics, 25(2), pp. 165-172.

a vector of the same length as the input, with outliers masked (or, if return_scores is true, the test statistics of the top max_outliers observations)

skinnider/modern documentation built on Feb. 20, 2020, 1:52 p.m.