Outliers can be detected by algorithms used for predictions. To illustrate, we use the k-nearest neighbor (kNN) clustering algorithm.