Skip to content

Outskewer: Using Skewness to Spot Outliers in Samples and Time Series

Notifications You must be signed in to change notification settings

sheymann/outskewer

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

9 Commits
 
 

Repository files navigation

Outskewer is a method to detect multiple outliers in data sets of numbers and in time series.

Outskewer uses a novel approach based on the notion of skewness (a measure of the distribution symmetry) and its evolution when extremal values are removed one by one. It is easy to interpret because values are classified as outliers, potential outliers or not outliers. The class of all values is unknown when the notion of outlier is not relevant in the considered data set.

###Key benefits

  • Easy to use because it requires no prior knowledge on the data,
  • the only parameter is the size of the time window for time series,
  • it may be used on-line.

###Reference S. Heymann, M. Latapy, and C. Magnien. "Outskewer: Using Skewness to Spot Outliers in Samples and Time Series", submitted to ASONAM'12.

###Implementations

###Want to discuss about the method? Fill an issue ticket to ask questions and discuss its behavior. I prefer to use github issue tracker than email for this kind of discussions, because this is public so it could benefit to anyone.

###Want to contribute? We need more implementations in other languages. An implementation in C is the priority.

About

Outskewer: Using Skewness to Spot Outliers in Samples and Time Series

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published