Machine Learning

Introduction to Machine Learning

Machine Learning (ML) is a branch of Artificial Intelligence that enables systems to learn from data and improve performance without being explicitly programmed.

Definition: Machine Learning is the study of algorithms that improve automatically through experience.

Why Machine Learning?

• Handles large volumes of data
• Learns patterns automatically
• Supports prediction and decision-making

ML vs Traditional Programming

Traditional Programming	Machine Learning
Rules + Data → Output	Data + Output → Rules

Types of Machine Learning

(a) Supervised Learning

• Learning with labeled data
• Input-output pairs are known

Examples: Classification, Regression

(b) Unsupervised Learning

• Learning with unlabeled data
• Finds hidden patterns

Examples: Clustering, Association

(c) Reinforcement Learning

• Learning through rewards and penalties
• Agent interacts with environment

Types Summary Table

Type	Data	Example
Supervised	Labeled	Spam detection
Unsupervised	Unlabeled	Customer segmentation
Reinforcement	Feedback	Game playing

Application Areas of Machine Learning

Area	Application
Healthcare	Disease prediction
Finance	Fraud detection
Marketing	Recommendation systems
NLP	Language translation
Robotics	Autonomous control

Decision Trees

A Decision Tree is a tree-structured model used for classification and regression.

Components

Component	Description
Root Node	Top decision point
Internal Node	Condition test
Leaf Node	Output decision

Simple Decision Tree Diagram

       Age?
      /    \
   Young   Old
    |       |
   Yes     No

Advantages & Disadvantages

Advantages	Disadvantages
Easy to understand	Overfitting
No data normalization	Sensitive to noise

Statistical Learning Methods

Statistical learning uses probability and statistics to model relationships in data.

Common Methods

Method	Use
Linear Regression	Prediction
Logistic Regression	Classification
k-Nearest Neighbor	Similarity-based
Support Vector Machine	Margin-based

Learning with Complete Data 

Learning with complete data means all variables and labels are fully observed.

Characteristics

• No missing values
• Direct probability estimation
• Easier model training

Naïve Bayes Model

Naïve Bayes is a probabilistic classifier based on Bayes’ theorem.

Bayes’ Theorem

P(C|X) = (P(X|C)P(C)) / P(X)

Assumption

• Features are conditionally independent

Example: Spam classification using word probabilities

Advantages & Limitations

Advantages	Limitations
Fast & simple	Independence assumption
Works with small data	Lower accuracy sometimes

Learning with Hidden Data

Hidden data refers to unobserved or missing variables during learning.

Challenges

• Incomplete information
• Indirect observation

Example

• Hidden customer intent
• Hidden system states

EM Algorithm (Expectation–Maximization)

EM algorithm is used to learn parameters when data is incomplete or hidden.

Steps of EM Algorithm

Step	Description
E-Step	Estimate missing data
M-Step	Maximize likelihood

EM Flow Diagram

Initial Guess
     ↓
  E-Step
     ↓
  M-Step
     ↓
 Converge

Applications

• Clustering
• HMM training
• Missing data problems

Reinforcement Learning

Reinforcement Learning (RL) is a learning method where an agent learns by interacting with environment.

Key Elements

Element	Description
Agent	Learner
Environment	World
Action	Decision
Reward	Feedback
Policy	Strategy

Reinforcement Learning Diagram

Agent → Action → Environment
  ↑                ↓
 Reward ← State ←——

Applications

• Robotics
• Game AI
• Self-driving cars

Comparison Summary

Learning Type	Data Type	Example
Supervised	Complete	Naïve Bayes
Unsupervised	Hidden	EM Algorithm
Reinforcement	Feedback	Q-learning

MCA Exam-Oriented Tips

• Write Bayes’ theorem clearly
• Draw EM and RL diagrams
• Compare supervised vs reinforcement learning
• Explain assumptions of Naïve Bayes
• Use tables for advantages & disadvantages

---

End of Notes