Enterprise Internet of Things Handbook

Unsupervised learning algorithms

An unsupervised ML algorithm is where we let the system figure out how the data has been structured or grouped, and based on this model, new incoming data can be placed in one of the clusters that define the data. For instance, say that a bank has a user's credit card usage pattern, and they build a cluster model of this pattern. When a new transaction takes place on that user's card, then based on the data model, the system can tell if this was a genuine purchase made by the intended user or if it is fraud.