Lecture 27 - Modeling and Evaluation 1

Occam's Razor, Overfitting, Data Splits

Announcements:

• Lab 4 graded; lab 5 on the way
• FP proposal due Friday night

Goals

• Understand how Occam's Razor applies to mathematical modeling and prediction tasks
• Know what it means for a model to overfit or underfit to its training data
• Understand why it's important to use separate validation and testing sets to evaluate a model
• Know what is meant by bias and variance in data-driven modeling
• Know how cross-validation works and why you might want to use it.

Occam's Razor

The simplest explanation for the data is the best explanation for the data.

Related: "Everything should be made as simple as possible, and no simpler." - Albert Einstein, maybe

Today turns out to be an all-handwritten day! Check out the written notes.