Supervised Learning

Supervised learning is the most widely used form of machine learning, and it underpins many of the AI tools people encounter every day. The model is trained on a dataset where every example already has a known, correct answer attached to it. The model studies those examples, spots patterns, and uses them to make predictions on data it has never seen before.

The name comes from the idea that the labelled dataset acts like a teacher. During training, the model makes a prediction, compares it against the known correct answer, and adjusts its internal settings to do better next time. Repeat this process across thousands or millions of examples, and the model becomes reliably accurate.

How it works#

The process starts with training data: a collection of input-output pairs. For each input, the label is the correct output the model should learn to predict. During training, the model passes each input through its calculations and produces a predicted answer. The difference between its prediction and the true label is called the loss. The model is adjusted, step by step, to reduce this loss.

Once training is complete, the model is tested on a separate set of examples it has never seen. If it performs well there, it can be trusted to work on genuinely new data in the real world.

Supervised learning problems fall into two main types. Classification tasks ask the model to assign an input to one of several categories, such as deciding whether a transaction is fraudulent or legitimate. Regression tasks ask the model to predict a numerical value, such as the expected energy consumption of a building over the next month.

Examples#

Email spam filtering. A spam filter is trained on thousands of emails already labelled as "spam" or "not spam". The model learns signals such as particular words, sender patterns, and formatting quirks that distinguish junk mail from genuine messages. When a new email arrives, the model predicts which category it belongs to and routes it accordingly.

Medical image analysis. Diagnostic AI tools used in hospitals are trained on thousands of medical scans, each labelled by a clinician with a diagnosis or a finding. The model learns to recognise visual patterns associated with conditions such as tumours or fractures. In practice, these tools assist clinicians rather than replacing them, flagging images that warrant closer attention.

Why it matters for business#

Supervised learning is the workhorse behind most commercial AI applications. If you have historical data with known outcomes, such as past sales figures, customer churn records, or quality-control results, supervised learning can often be used to build a model that predicts future outcomes from that data. The main requirement is having enough labelled examples and the means to check the model's predictions before relying on them.

Understanding how supervised learning works helps you ask the right questions when evaluating AI tools: what was the model trained on, how large was the dataset, and how was its accuracy measured? The tutorial AI vs ML vs Deep Learning explains where supervised learning sits among the other approaches, and our /paths page offers guided learning routes.