Exploring use cases of anomaly detection

Before moving on to some specific algorithms for anomaly detection, let's first consider some use cases that are often done with anomaly detection.

Fraud detection in financial institutions

A very common use case for anomaly detection is the detection of fraud in financial institutions. Banks generally have a lot of data, as almost everyone has one or more bank accounts that are used on a regular basis. All these usages generate a huge amount of data that can help banks to improve their services and their profits. Fraud detection is a key component of data science applications in banks, together with many other use cases.

A common use case for fraud detection is to automatically detect credit card fraud. Imagine that your card or card details have been stolen and someone is fraudulently using them. This leads to fraudulent transactions, which could be automatically detected by a machine learning algorithm. The bank could then automatically...