Chapter 8

Question 1

What is supervised learning?

Answer 1

A way to predict a value, y, from obeservations x.

Question 2

What does it mean that the function is linear?

Answer 2

1) additivity (f(x+y) = f(x) + f(y)

2) Homogeneity (f(ax) = af(x))

Question 3

How to maximize w in linear regression?

Answer 3

Minimize for error function

Question 4

What is logistic regression?

Answer 4

A binary variable can be described with a bernoulli distribution.

The bernoulli distribition can be expressed via a function (sigma(z)) that is the logistic sigmoid.

Long story short: we can re-write binary expectations via probvabilities so we can talk about the probability of a given variable beiong of either class.