Using K-nearest neighbors to find outliers

Machine learning tools can help us identify observations that are unlike others when we have unlabeled data – that is, when there is no target or dependent variable. Even when selecting targets and features is relatively straightforward, it might be helpful to identify outliers without making any assumptions about relationships between features, or the distribution of features.

Although we typically use K-nearest neighbors (KNN) with labeled data, for classification or regression problems, we can use it to identify anomalous observations. These are observations where there is the greatest difference between their values and their nearest neighbors' values. KNN is a very popular algorithm because it is intuitive, makes few assumptions about the structure of the data, and is quite flexible. The main disadvantage of KNN is that it is not as efficient as many other approaches, particularly parametric techniques such as linear regression...