Lecture 13 – Simple Linear Regression
Presented by Anthony D. Joseph and Suraj Rampure
Content by Suraj Rampure and Ani Adhikari
A reminder – the right column of the table below contains Quick Checks. These are not required but suggested to help you check your understanding.
Introduction and recap of the modeling process.
The correlation coefficient and its properties.
Defining the simple linear regression model, our first model with two parameters and an input variable. Motivating linear regression with the graph of averages.
Using calculus to derive the optimal model parameters for the simple linear regression model, when we choose squared loss as our loss function.
Visualizing and interpreting loss surface of the SLR model.
Interpreting the slope of the simple linear model.
Defining key terminology in the regression context. Expanding the simple linear model to include any number of features.
RMSE as a metric of accuracy. Multiple R-squared as a metric of explained variation. Summary.