COMP 5013 at Lakehead, built from the course textbook (An Introduction to Statistical Learning with Python): regression, classification, resampling, regularization, splines and GAMs, trees, SVMs, neural networks, PCA and clustering, plus the Python used in the labs and the mixture and ICA material from the course outline.
You’re reading a free preview. Sign up to see every topic and question — and to play this deck as games.
ISLP section 2.1: the model Y = f(X) + ε, why we estimate f (prediction or inference), parametric and non-parametric methods, the flexibility and interpretability trade-off, and the main kinds of learning problem.
35 total questions
Card #1
In ISLP's general model Y = f(X) + ε, what does f stand for?
Card #2
What two assumptions does ISLP make about the error term ε in Y = f(X) + ε?
Card #3
Reducible error
1. A company wants to identify which people are likely to respond positively to a mailing, using demographic variables. It does not care how each variable relates to the response. Which kind of modelling is this?
2. A real estate analyst asks: "How much extra is a house worth if it has a view of the river?" In ISLP's terms this is:
3. You only want to predict tomorrow's stock price as accurately as possible, and interpretability does not matter. What does ISLP say about simply using the most flexible model available?
1. When the goal is prediction, f is often treated as a ___: we do not care about its exact form as long as its predictions are accurate.
Hint: Two words: something you cannot see inside
2. The irreducible error provides an ___ bound on the accuracy of our prediction for Y, and in practice that bound is almost always unknown.
Hint: The opposite of lower
3. β0, β1, ..., βp in a linear model are most commonly estimated by (ordinary) ___.
Hint: Covered in Chapter 3
