Machine Learning (ML) is a branch of Artificial Intelligence in which machines learn patterns from data or experience and use that learning to make predictions, decisions, or discover useful information.
Types of Machine Learning #
Machine Learning can be broadly divided into four important learning approaches:
1. Supervised Learning #
Simple Definition #
You give the machine some data along with the correct answers, and the machine learns to predict answers on new, unseen data.
Formal Definition #
Supervised Learning is a type of Machine Learning in which a model learns from labeled data, where input features and their corresponding target/output values are provided during training. The trained model is then used to predict outputs for new or unseen data.
Supervised Learning Diagram #
Example #
| IQ | CGPA | Placement |
|---|---|---|
| 87 | 7.1 | Yes |
| 75 | 6.3 | No |
| 111 | 8.9 | Yes |
| 75 | 7.3 | ? |
Here, IQ and CGPA are input features and Placement is the target/output. The model learns from the labeled examples and predicts the placement for a new student.
Types of Supervised Learning #
Regression #
Regression is a supervised learning technique where the target/output variable is numerical or continuous.
- House price → ₹50 lakh
- Salary → ₹8 lakh
- Temperature → 32.5°C
- CGPA → 8.5
Remember: Numerical Target → Regression
Classification #
Classification is a supervised learning technique where the target/output belongs to a category or class.
- Placement → Yes / No
- Email → Spam / Not Spam
- Student → Pass / Fail
- Disease → Positive / Negative
Remember: Categorical Target → Classification
2. Unsupervised Learning #
Simple Definition #
You give the machine data without correct answers, and the machine tries to find patterns, groups, or relationships in the data by itself.
Formal Definition #
Unsupervised Learning is a Machine Learning approach in which a model learns from unlabeled data without predefined target outputs and attempts to discover hidden patterns, structures, or relationships.
Unsupervised Learning Diagram #
Example #
Suppose an e-commerce company has customer data such as age, income, purchase frequency, and total spending. However, there are no predefined labels such as “Premium Customer” or “Regular Customer”.
A clustering algorithm can discover groups of customers with similar characteristics.
Common Unsupervised Learning Techniques #
- Clustering – Groups similar data points.
- Association – Finds relationships between items.
- Dimensionality Reduction – Reduces the number of features.
3. Semi-Supervised Learning #
Simple Definition #
You give the machine a small amount of data with correct answers and a large amount of data without answers, and the machine learns from both.
Formal Definition #
Semi-Supervised Learning is a Machine Learning approach that uses a combination of a small amount of labeled data and a large amount of unlabeled data to train a model.
Semi-Supervised Learning Diagram #
Example #
Imagine you have 1,000 images, but only 100 images are labeled. The remaining 900 images have no labels. Labeling all images manually can be expensive and time-consuming. Semi-supervised learning can use the small labeled dataset together with the larger unlabeled dataset.
4. Reinforcement Learning #
Simple Definition #
The machine learns by interacting with an environment, taking actions, and receiving rewards or penalties based on those actions.
Formal Definition #
Reinforcement Learning (RL) is a Machine Learning approach in which an agent interacts with an environment, takes actions, and learns a policy through rewards or penalties. The objective is to maximize cumulative reward over time.
Reinforcement Learning Diagram #
Important Components #
- Agent: The learner or decision-maker.
- Environment: The world in which the agent operates.
- State: The current situation of the agent.
- Action: An operation performed by the agent.
- Reward: Feedback received after taking an action.
- Policy: A strategy that determines which action to take.
Example #
Consider a game-playing AI. The AI is the agent and the game is the environment. The AI chooses an action, such as moving left or right. If the action helps the AI progress, it may receive a positive reward. If the action leads to failure, it may receive a negative reward or penalty.
5. Comparison of Machine Learning Approaches #
| Learning Type | Data / Feedback | Main Goal | Examples |
|---|---|---|---|
| Supervised | Labeled data | Predict target | Classification, Regression |
| Unsupervised | Unlabeled data | Find patterns / structure | Clustering, PCA |
| Semi-Supervised | Small labeled + large unlabeled data | Learn using both | Image / Text Classification |
| Reinforcement | Rewards / penalties | Learn actions | Games, Robotics, Control |
6. Easy Way to Remember #
Supervised: Data + Correct Answer → Learn → Predict
Unsupervised: Data without Answer → Find Patterns
Semi-Supervised: Small Labeled + Large Unlabeled → Learn
Reinforcement: Action → Reward/Penalty → Learn
7. One-Line Definitions for Exams #
Supervised Learning: You give the machine some data along with the correct answers, and the machine learns to predict answers on new, unseen data.
Unsupervised Learning: A Machine Learning approach in which the model learns from unlabeled data to discover hidden patterns, structures, or groups.
Semi-Supervised Learning: A Machine Learning approach that uses a small amount of labeled data together with a large amount of unlabeled data for learning.
Reinforcement Learning: A Machine Learning approach in which an agent learns by interacting with an environment and receiving rewards or penalties for its actions.
Conclusion #
The four learning approaches differ mainly in how the machine receives information during learning:
- Supervised Learning → Learn from correct answers.
- Unsupervised Learning → Discover patterns without correct answers.
- Semi-Supervised Learning → Learn from both labeled and unlabeled data.
- Reinforcement Learning → Learn through actions and rewards.
