Chapter 14: Outlier Detection Using Unsupervised Machine Learning
In Chapter 8, Outlier Detection Using Statistical Methods, you explored parametric and non-parametric statistical techniques to spot potential outliers. The methods were simple, interpretable, and yet quite effective.
Outlier detection is not straightforward, mainly due to the ambiguity surrounding the definition of what an outlier is specific to your data or the problem that you are trying to solve. For example, though common, some of the thresholds used in Chapter 8, Outlier Detection Using Statistical Methods, are still arbitrary and not a rule that you should follow. Therefore, having domain knowledge is vital to making the proper judgment when spotting outliers.
In this chapter, you will be introduced to a handful of machine learning-based methods for outlier detection. Most of the machine learning techniques for outlier detection are considered unsupervised outlier detection methods, such as Isolation Forests...
