Part 2.

Part 2 covers the main algorithms and libraries for outlier detection in Python.

In chapter 5 we introduce outlier detection with scikit-learn. In chapter 6 we introduce the PyOD library, which is probably the most comprehensive library available for outlier detection for numeric tabular data in Python. In both chapters 5 and 6, we cover the algorithms provided (each of these libraries provides several tools for outlier detection), explain how they work, and describe how they may be used in your projects.

In chapter 7 we describe several other algorithms, tools, and libraries. These are also very useful and effective but are more difficult to find than those in scikit-learn or PyOD. They will help you understand outlier detection itself better (examining them provides a fuller understanding of the breadth of approaches available to identify outliers), help you see how to develop your own detectors where necessary, and provide a set of tools that are useful in themselves. They also include some detectors suited for categorical data, which is not directly supported by scikit-learn or PyOD.