Machine Learning

A. Supervised Learning

In supervised learning, the model learns using labeled data, where the correct answer is already known.

The goal is to learn the relationship between input and output.

Example

Suppose you have student marks and know whether they passed.

The model learns:

More study hours usually lead to passing.

Later, if a student studies 6 hours, the model predicts "Pass."

Applications

1. House price prediction
2. Email spam detection
3. Disease prediction
4. Credit approval

B. Unsupervised Learning

Here, the data has no labels.

The algorithm tries to discover hidden structures or patterns on its own.

Example

A supermarket has customer purchase history but no predefined customer categories.

The algorithm groups customers into clusters such as:

1. Frequent buyers
2. Occasional buyers
3. Premium customers

No one tells the algorithm these groups in advance—it discovers them.

Applications

1. Customer segmentation
2. Market basket analysis
3. Recommendation systems

C. Reinforcement Learning

The model learns by interacting with an environment and receiving rewards or penalties based on its actions.

The goal is to maximize the total reward over time.

Example

A robot learns to walk:

1. Correct movement → Reward
2. Incorrect movement → Penalty

After many attempts, it learns the best walking strategy.

Applications

1. Self-driving cars
2. Robotics
3. Chess
4. Video games