jamesalday/doublemadoutliers

Uses the Double Median Absolute Deviation test to find outliers in data

Maintainers

Details

github.com/JamesAlday/DoubleMadOutliers

Homepage

Source

Issues

Installs: 15

Dependents: 0

Suggesters: 0

Security: 0

Stars: 3

Watchers: 0

Forks: 1

Open Issues: 0

dev-master 2018-06-20 19:59 UTC

Requires

  • php: ^5.3.3 || ^7.0

Requires (Dev)

MIT 11934bf5add2cb001db9cdffe79ae609e9ed931a

mathdeviationstatitics

This package is not auto-updated.

Last update: 2024-06-09 03:24:09 UTC

README

This is a small library class that tests asymmetric data for outliers based on the 'Double MAD' method.

The data is split into 2 'legs' around the median, and each leg is tested for values that are larger than the median absolute deviation for that leg. This prevents high/low outliers in the data from canceling each other out or smaller outliers being hidden in widely varied data.

This library is an adaptation of the algorithm/R script written by Peter Rosenmai on Eureka Statistics. He gives a much more in-depth analysis of the reasoning behind this method, and it well worth a read if you intend to use this class.

https://eurekastatistics.com/using-the-median-absolute-deviation-to-find-outliers/