The main difference between 'supervised' and 'unsupervised' learning is that:

Supervised learning uses labelled training data (features + target) for Classification (predict category — fraud/not fraud, default/no default) and Regression (predict continuous value — loan amount, house price). Algorithms of Supervised Learning are Linear/Logistic Regression, Decision Trees, SVM, Random Forest, Neural Networks. Unsupervised learning uses data with no labels for Clustering (K-Means, DBSCAN — customer segmentation), Dimensionality Reduction (PCA — compressing features) and Association Rules (Apriori — product bundling). In Banking, Supervised learning is used for credit scoring and Unsupervised learning is used for detecting unusual transaction clusters (potential fraud rings).