Lecture 3


GLM tie-up; re-orientation

Fitting a model function involves separate consideration of the model function and conditional distributions:

See Lecture 4 for a better version of this table

Model function assumption? Distributional Assumption? Method
no no kNN, loess, random forests, other “machine learning” techniques
yes no least squares (example: linear regression)
yes yes MLE (example: GLM, including linear regression)
no yes Machine learning techniques, with likelihood loss function (which often amounts to least squares, as usual, when estimating mean)

Lecture 2:

This week:

Learning Goals

By the end of this lecture, students are expected to be able to: