Clustering to categorize transactions into N groups based on feature similarity.

Unsupervised learning to identify transactions likely to be fraudulent.

Supervised learning to forecast the location of a transaction.

Supervised learning to detect fraudulent transactions.

Reinforcement learning to estimate the location of a transaction.

Unsupervised learning to estimate the location of a transaction.