Why do we use the sigmoid function at the end?

Course Q&A Machine Learning Specialization Supervised ML: Regression and Classification
1

Hi!
This course is amazing so far - learning so much!
I’m beginning looking at logistic regression, and we are told that f(x) = g(w.x + b),
in other words to put our answer through the sigmoid function at the end.

My question is, why can’t we put all of our variables through the sigmoid function before we use linear regression, in a similar way to how we created new features for polynomial regression?

Thanks for any help!

2

The sigmoid function re-scales the value of (w.x + b) so it is in the range of 0 to 1.

3

That makes sense, thank you!