Introduction to Machine Learning
Course: AI & Machine Learning Foundations
PREVIOUS COURSES
Regression – How Machines Predict Numbers
Today, we will learn how computers can predict numerical values from data.
This lecture introduces the basic ideas behind Machine Learning before studying
our first prediction algorithms.
By the end of today's session, you will understand:
• Why Machine Learning is needed.
• How Machine Learning is different from traditional programming.
• The complete Machine Learning workflow.
• The main types of Machine Learning.
• The basic idea of Regression.
• The first Regression algorithms used in AI.
KEY NOTE
Today is the beginning of real Machine Learning. Everything you learned in previous courses will now be used together.
Learning Objectives
After completing today's lecture, you should be able to:
• Explain what Machine Learning is.
• Differentiate between Supervised, Unsupervised, and Reinforcement Learning.
• Describe why Machine Learning is important.
• Compare Traditional Programming and Machine Learning. • Understand the idea of Regression.
• Identify situations where Machine Learning should be used. • Recognize common Regression algorithms.
• Explain the Machine Learning Pipeline. • Select suitable evaluation metrics for Regression problems.
KEY NOTE
Today's goal is not only to learn algorithms, but also to learn how Machine Learning engineers think.
Lecture Roadmap
Today's lecture is organized into the following sections:
Part Topic
Part 1 Introduction to Machine Learning
Part 2 Regression Fundamentals
Part 3 Linear Regression
Part 4 Polynomial Regression
Part 5 Regularization
Part 6 Decision Tree Regressor
Part 7 Regression Metrics
Part 8 Final Comparison
Each section builds on the previous one.
Do not think of them as separate topics. They form one complete Machine Learning story.
Intro to ML Regression Fund. Linear Reg. Polynomial Reg. Regularization Decision Tree Metrics Comparison
Before We Start...
Before learning any algorithm, ask yourself one important question:
Why do we need Machine Learning at all?
If traditional programming can already solve many problems, why did Machine Learning become one of the fastest-growing fields in AI?
Understanding this answer will make every algorithm we study much easier to understand.
ACTIVITY / DISCUSSION QUESTION
Discuss with your partner:
Can you think of a problem that would be difficult to solve using only fixed programming rules?
KEY NOTE
Always understand the problem before learning the solution.
What is Machine Learning?
Machine Learning ML) is a branch of Artificial Intelligence. WORKFLOW VISUALIZED
Instead of writing every rule manually, we allow the computer to learn patterns from
data.
The computer studies examples.
Then it uses what it learned to make predictions on new data.
Machine Learning does not memorize every example.
It tries to discover the hidden relationships inside the data.
SIMPLE DEFINITION
Machine Learning is the process of learning patterns from data to make
predictions or decisions.
AI Connection: Many modern AI systems, including recommendation systems, spam filters, fraud detection, and ChatGPT, rely on Machine Learning techniques.
Traditional Programming vs Machine Learning
Traditional Programming follows fixed rules written by a programmer.
Machine Learning learns the rules from examples.
Traditional Programming Machine Learning
Inputs + Rules → Output Inputs + Outputs → Learned Model
Rules are written manually Rules are learned automatically
TRADITIONAL PROGRAMMING
Good for fixed tasks Good for complex patterns
Input Rules Output
MACHINE LEARNING
Input + Learning Model →
Correct Output Algorithm Prediction
EXAMPLE TRADITIONAL
Calculator
EXAMPLE ML
House Price Prediction
KEY NOTE
Machine Learning is not replacing programming. Programming is still needed to build Machine Learning systems.
The Machine Learning Pipeline
Building a Machine Learning model follows a structured workflow.
PROFESSIONAL PROCESS WORKFLOW
Each stage has a specific purpose.
The typical Machine Learning Pipeline is:
• Define the Business Problem
• Collect Data
• Prepare the Data
• Train the Model
• Evaluate the Model
• Deploy the Model
• Monitor and Improve the Model
Skipping one stage usually reduces the quality of the final system.
KEY NOTE
Machine Learning is a complete process, not only training an algorithm.
Understanding Each Pipeline Stage
Stage Purpose
Business Problem Define the objective
Collect Data Gather useful information
Prepare Data Clean and organize data
Train Model Learn patterns
Evaluate Measure performance
Deploy Use the model in real applications
Monitor Improve the model over time
Every successful AI project follows these steps.
Some projects repeat the pipeline many times until the model reaches the desired performance .
AI Connection: Companies such as Google, Microsoft, and OpenAI continuously retrain and improve their Machine Learning models.