Supervised Learning Process

Introduction to Linear Regression

• Focus: Supervised learning process and linear regression model
• Linear regression: Fitting a straight line to data, widely used
• Concepts apply to other machine learning models

Example Problem: Predicting House Prices

• Task: Predict the price of a house based on size
• Data: House sizes and prices from Portland, USA

Plotting Data Points

• Horizontal axis: Size of the house (sq ft)
• Vertical axis: Price of the house (thousands of dollars)
• Each data point: House with a size and a price

Use Case

• Real estate scenario: Predict price for a 1250 sq ft house
• Use linear regression to estimate the price: ~$220,000

Supervised Learning Model

• Training model with data that has right answers
• Example: House size (input) and price (output)
• Linear regression: Predicts numbers (regression problem)

Types of Supervised Learning

• Regression Model: Predicts numbers (e.g., house prices)
  • Linear regression, other regression models
• Classification Model: Predicts categories
  • Examples: Cat vs Dog, medical condition prediction

Differences Between Classification and Regression

• Classification: Finite set of possible outputs
• Regression: Infinite number of possible outputs

Data Visualization

• Plot on graph and as data table
• Data Table: Inputs (size) and outputs (price)
• Horizontal and vertical axes correspond to columns in data table

Example Data Point

• House size: 2104 sq ft
• Price: $400,000

Notation in Machine Learning

• Training set: Data used to train model
• Input variable (feature): x, e.g., size of the house
• Output variable (target): y, e.g., price of the house

Notation Examples

• First example: x = 2104, y = 400
• Total training examples: m = 47
• Single training example: (x, y) = (2104, 400)
• Using i-th example notation: (x^(i), y^(i))
• Superscript i: Index of training example, not exponentiation

Conclusion

• Covered: Training set and associated notation
• Next: Feeding training set to learning algorithm