Supervised Learning

What Supervised Learning is

In supervised learning every training example carries a ground-truth answer. The model's job is to generalise the mapping from inputs to those answers rather than memorise the examples.

How it works

Data is labelled by people or by an existing process, split into train and test sets, and fed to an algorithm — linear models, gradient-boosted trees or neural networks — that minimises prediction error. Accuracy, precision, recall and calibration are measured on held-out data.

Why it matters

It remains the workhorse of applied machine learning because business problems usually come with historical outcomes attached: did the customer churn, was the transaction fraudulent, did the part fail.

Common uses

• →Email spam filtering
• →Loan default prediction
• →Medical diagnosis support
• →Sentiment classification