OutlierDetection.jl: A modular outlier detection ecosystem for the Julia programming language

TL;DR

OutlierDetection.jl is a comprehensive, high-performance, and modular ecosystem for outlier detection in Julia, enabling scalable algorithms, standardized interfaces, and best practices for development.

Contribution

It introduces a flexible, scalable outlier detection ecosystem in Julia with standardized interfaces and support for model composition, unlike previous packages.

Findings

Provides a range of high-performance outlier detection algorithms.

Enables development of highly-scalable algorithms using Julia.

Implements best practices like unit testing and continuous integration.

Abstract

OutlierDetection.jl is an open-source ecosystem for outlier detection in Julia. It provides a range of high-performance outlier detection algorithms implemented directly in Julia. In contrast to previous packages, our ecosystem enables the development highly-scalable outlier detection algorithms using a high-level programming language. Additionally, it provides a standardized, yet flexible, interface for future outlier detection algorithms and allows for model composition unseen in previous packages. Best practices such as unit testing, continuous integration, and code coverage reporting are enforced across the ecosystem. The most recent version of OutlierDetection.jl is available at https://github.com/OutlierDetectionJL/OutlierDetection.jl.