CIS 421/521 - Logistic Regression and Stochastic Gradient Descent - 2020-11-19 - Shared screen with speaker view
what if we have 0 prob for y = 1, then we will have log(0)?
Is 1.17 when we’re predicting y = 1 or y = 0?
oy thank you for dealing with that all!!
How come the curve goes up after we cross wmin?
The curve is arbitrary
We were trying to find the min of the curve
Why can’t we find this min using mathematical ways, eg. Taking the derivative = 0?
Is it possible to instead find those critical points and end points, and simply compute f(x) for all of them to find the global minimum? Or would this process usually take longer than taking these learning steps
Hi professor, just want to confirm that we have ten late days in total for the semester right