dailybrief

Foundations · 1 min read · Updated 2026-10-05

Common machine learning methods

Four common machine learning methodsA line through dots, a yes/no decision tree, coloured groups, and a small neural network.Line fittingPredicts a numberDecision treeAsks yes/no questionsClusteringFinds groupsNeural networkLearns tricky patternsRain?UmbrellaSunhatyesno
Four popular methods and what each is good at.

Line fitting (linear regression)

It draws the best straight line through dots to predict a number. For example, a bigger house usually costs more, so the line goes up.

Decision tree

It asks a chain of yes or no questions, like 'Is it raining?' Each answer leads to the next question until it reaches a decision.

Clustering

It puts similar things in the same group without being told the groups. A popular version is called k-means.

Neural network

Many tiny connected units that learn patterns too tricky for simple rules, like faces, speech and language. This is the method behind modern chatbots.

Key takeaways

  • Different problems need different methods.
  • Simple methods are fast and easy to explain.
  • Neural networks handle the hardest patterns.

Quick questions

Which method is best?

It depends on your data and your goal. Simple methods are often enough.

What is an algorithm?

A step-by-step recipe that a computer follows.

Do chatbots use these?

They mainly use neural networks. The other methods are still common elsewhere.

Keep learning

Types of machine learning at a glanceMachines learn in three main ways: from examples that come with answers (supervised), by finding groups on their own…1 min readRead →Deep learning and neural networksA neural network is a computer program made of many small connected parts, a bit like brain cells. It learns by slowly…1 min readRead →