Unsupervised Machine Learning Approaches for Outlier Detection in Time Series, using Python

In this post, I cover some of my favorite methods for detecting outliers in time series data. There are many different approaches for detecting anomalous data points; for the sake of brevity, I only focus on unsupervised machine learning methods in this post. I deal with time series data a lot, and it's not uncommon for data sets to experience unexpected drops or spikes, flat lines, or phase shifts. Each of these situations qualifies as an'anomaly' -- something out of the ordinary when compared to the behavior of the sequence as a whole. Detecting anomalies in a time series is important for a variety of reasons.