8.6 Supervised Learning Practical Examples

Putting supervised learning to work follows a repeatable workflow: collect and label data, split it into training and test sets, train a model on the training set, then evaluate on the held-out test set to estimate real-world performance. Choosing the right model depends on whether the target is a category (classification) or a number (regression), and on needs like accuracy versus interpretability.

Consider a mobile-money provider in Ethiopia: it can use classification to flag suspicious transactions as fraud or not, regression to forecast a customer's monthly spending, a random forest for robust fraud accuracy, and KNN to recommend offers similar to those liked by nearby customers.