DEV Community

Cover image for Types of Machine Learning - Explained Simply
Jasim Alam
Jasim Alam

Posted on

Types of Machine Learning - Explained Simply

Supervised Learning : Learning with a teacher

Imagine a student learning from past exams.
The machine gets both the questions and the correct answers.
Over time, it learns the pattern.

Real-Life Examples:

  • Spam Email Detection
  • House Price Prediction
  • Disease Diagnosis
  • Stock Market Forecasting
  • Weather Predictions

Example:

You show a computer 1000 labeled images:

  • dog
  • cat

Later, it can say, “That’s a cat!” when shown a new image.


Unsupervised Learning : Learning without a teacher

Like a child sorting toys by shape or color without knowing their names.

The machine explores data without labels and discovers hidden patterns.

Real-Life Examples:

  • Customer Segmentation for targeted ads
  • Market Basket Analysis (product recommendations)
  • Fraud and Anomaly Detection

Example:

Give it 1000 fruit pictures — no names. It might group them like:

  • Round and red fruits
  • Long and yellow fruits
  • Small and purple fruits

It doesn’t know the fruit names but it finds structure.


Reinforcement Learning : Learning by doing

Like training a pet with rewards and corrections.

The machine (called an agent) tries different actions, gets rewards or penalties, and learns what works best over time.

Real-Life Examples:

  • Self-driving Cars
  • Game AI (Chess, Go, video games)
  • Robotics and Automation
  • Stock Trading Bots

Example:

  • A robot tries to walk.
  • Falls result in no reward.
  • Moves forward result in a reward.
  • It learns to stay balanced and move smart.

Top comments (0)