Logistic Regression

Question 1

What does LR predict?

Answer 1

LR predicts if sth is true or false (like mices are obese or they are not).

Question 2

What does the LR function looks like graphically?

Answer 2

An S shape, goes from 0 to 1

Question 3

How are probabilities relevant in classification in LR?

Answer 3

It is a classification algorithm with output values between 0-1

If the probability that a mice is obese is 0.51(so in the middle of the S shape graph), then it gets classified as obese, vice versa.

Question 4

Can LR be filled with discrete or continous variables?

Answer 4

Both

Question 5

What does maximum likelihood do?

Answer 5

Doesn´t use Least Squared method or R**2, but maximum likelihood.

You have all your points, pe. sth that looks like a normal distribution. You want to find out where it is centered. You try to situate the bell left, then go further right, right and so on. At the end you take the spot where the likelihood to observe the data you have measured is the highest.

Question 6

How to calculate an odd?

Answer 6

Divide a probability by another probability. Coin flip. Heads / Tails= 0,5 / 0,5=1 So the odds are 1:1