Machine Learning course, Lesson 2

Types of Machine Learning

Learn when to use supervised, unsupervised and reinforcement learning. Compare regression, classification and clustering, and practise selecting an approach for real business problems. Work through eight concepts with interactive visuals, activities and quiz questions, then complete a customer-churn business challenge.

Lesson progress0%

Concept 1 of 80 of 8 concepts completed

The Three Main Learning Setups

Distinguish supervised, unsupervised and reinforcement learning by the feedback available to the learner.

Understand the concept

Machine Learning approaches differ in what information guides learning. Supervised learning uses examples with known outcomes, unsupervised learning looks for structure without supplied target labels, and reinforcement learning learns through actions and rewards. These are different learning setups, not a ranking from worst to best.

Business connection

A retailer can use labelled customer outcomes to predict churn, explore customer groups without existing labels, or use reward feedback to optimise a sequence of decisions. The business question and available data guide the choice.

Visual explanation

Explore three learning setups

Select each setup to inspect what the learner receives and the kind of problem it can address. No single setup is right for every task.

Supervised

The learner sees examples together with the correct outcome. Example: predict sales from historical sales data, or estimate churn from a labelled churn dataset.

Interactive activity

Match the setup to its feedback

Match each description with the learning setup that fits it best.

Choose one setup for each description.

A retailer wants to discover customer groups without pre-existing group labels.

An agent improves its actions using rewards from repeated interaction.

Examples include known sale prices that the model should estimate.

Machine Learning Basics for Business

This machine learning for beginners lesson explains the core ideas behind modern predictive tools. Instead of memorising definitions, you practise the same reasoning an analyst uses when a manager asks whether data can help with a decision.

What is machine learning?

Machine learning is a way of building software that learns patterns from examples. A machine learning model is the learned result of training, and it is useful only when its predictions are tested and used with care.

How does a machine learning model work?

A model takes input features, such as advertising spend or days since a customer last ordered, and produces an estimate of a target, such as sales or churn. Training data supplies the examples, and evaluation on unseen data shows whether the pattern holds for new cases.

Key terms

Supervised learning: Learning from examples with known target outcomes.
Regression: A supervised task that estimates a numerical value.
Classification: A supervised task that predicts a category or class.
Clustering: Grouping observations by similarity without supplied group labels.
Reinforcement learning: Learning actions through interaction and reward feedback.

Frequently asked questions

What are the three main types of machine learning?

The three common learning setups are supervised learning, unsupervised learning and reinforcement learning. They differ in the feedback available during learning.

What is supervised learning?

Supervised learning uses examples paired with known target outcomes to learn how inputs relate to a target.

What is the difference between regression and classification?

Regression estimates a numerical value, while classification predicts a category or class.

What is unsupervised learning used for?

Unsupervised learning can explore structure in data without supplied target labels, for example by grouping similar observations with clustering.

How does reinforcement learning work?

An agent interacts with an environment, takes actions and receives reward feedback that helps it learn a strategy over time.