Machine Learning - Coursera

Notes from the Coursera Machine Learning courses from University of Washington on Coursera.

Machine Learning Foundations

Week 2 - Regression: Predicting House Prices

  • Notes
    • Predicting house prices using sq ft
    • Regression models
    • Regression coefficients
  • Assignment 1
    • Explore house sale data using SFrame
Machine Learning: Regression

Week 1 - Simple Linear Regression

Week 2 - Multiple Regression

  • Notes
    • Multiple Regression model - single variable
    • Multiple Regression model - multiple variable
    • Coefficients of regression with multiple features
    • Matrix Arithmetic
    • Multiple Regression in matrix notation (single input)
    • Multiple Regression in matrix notation (multiple inputs)
    • Minimize Cost - Gradient Descent
  • Assignment 2.1
    • Explore house sale data using SFrame
    • Train vs Test data
    • Build multiple regression model using Square feet, number of bedrooms and number of bathrooms
    • Explore model
    • Interpretations of Coefficients
    • Making Predictions
    • RSS
    • Add more features
    • Comparing models
  • Assignment 2.2
    • Implement Gradient Descent Algorithm
    • Compare single variable vs multiple variable models
    • Compare with in built graphlot model

Week 3 - Assessing Performance

  • Notes
    • Access model performance
    • Training error
    • Generalization error
    • Test error
    • Sources of error -> irreducible error, bias and variance
    • Variance Bian tradeoff
  • Assignment 3
    • Use matplotlib to visualize polynomial regressions
    • Use a validation set to select a polynomial degree
    • Assess the final fit using test data
  • price vs sqft

    price vs quadratic function(sqft)

    price vs cubic function(sqft)

    price vs 15th polynomial function(sqft)