Complete Outlier Detection Algorithms A–Z
The companion repository for a Udemy course on outlier detection, with lecture notebooks plus reusable Python utilities, tests, exercises and benchmarks.
This is the official companion repository for a Udemy course on outlier detection, spanning statistical rules, classical machine learning, ensembles and autoencoders across three sections and eighteen lectures. It provides two learning tracks: lecture-aligned notebooks that follow the recorded course, and a modern reference track of reusable Python utilities, automated tests, exercises and current engineering guidance.
Working through the material teaches how to distinguish global, contextual and collective anomalies, the assumptions behind statistical, distance, density, isolation, kernel and reconstruction-based methods, how to implement univariate, multivariate and high-dimensional detection workflows without leaking information during feature scaling, how to evaluate rare-event models with suitable metrics, and how to compare algorithms with reproducible experiments.
It is distributed as Jupyter notebooks and a reusable Python package.