Part 1.

Part 1 covers the basic ideas of outlier detection: what outliers are, some techniques to find outliers in data, and how to manage outlier detection projects.

In chapter 1, we cover the idea of outliers and provide some examples of where outlier detection may be used, along with high-level descriptions of how outlier detection may be applied to these cases. We look at the subjective nature of outliers, provide some history of outlier detection, and describe the place of outlier detection in machine learning generally.

In chapter 2, we introduce techniques for outlier detection with simple statistical methods (such as z-score and interquartile range), which can find rare or extreme values in sequences of values. These techniques are straightforward but are often all that is necessary for outlier detection projects. Where more sophisticated techniques are necessary, these often build upon the ideas presented in this chapter.

In chapter 3, we extend the discussion from one-dimensional series of values to tables of data and introduce some of the more common approaches to outlier detection for tabular data, along with implementations in Python.

In chapter 4, we look at how outlier detection projects can execute from start to finish. This can sometimes be left out of the discussion with outlier detection but is necessary for an effective process.