Lecture 8 - Logistic Regression

Question 1

What does logistic regression do?

Answer 1

It takes a linear regression and applies the “logistic function” to the output to map them to a fixed range (0, 1) to provide a probability that the sample is the positive class.

Question 2

What do we do for training the cost function of a logistic regression function?

Answer 2

Cross-Entropy loss or log loss, measures the performance of a classification model that outputs a probability

Question 3

What shape does the logistic regression function take?

Answer 3

s-curve that maps the entire domain of x to into the range (0,1)

Question 4

What does the cross-entropy function do?

Answer 4

Cross-entropy loss increases as the predicted probability diverges from the actual label. Basically, the worse it gets the higher the penalty.