Recognizing different types of anomalies
Before we dive into Amazon Lookout for Metrics, I recommend that you read a few definitions first. If you need a refresher about what different types of anomalies look like in time series data, you can read the What are the different approaches to tackle anomaly detection? section at the beginning of Chapter 8, An Overview of Amazon Lookout for Equipment. This paragraph will give you an overview of the different types of anomalies we can find in time series data, along with a description of the different approaches we can use when building a custom outlier or anomaly detection system from scratch.
If you head over to the Amazon Lookout for Metrics home page (https://aws.amazon.com/lookout-for-metrics), you will read that it uses ML techniques to automatically detect anomalies that are defined as outliers from the norm. Amazon Lookout for Metrics looks for deviations in live or real-time univariate time series data. The keyword in this description...