3
Total questions
1
Awaiting your reply
Needs a reply
0
Sent, awaiting admin review
1
Answered
Delivered to student
Awaiting your reply
Please reply within 12–24 hours of admin approval
Sep 6 — Machine Learning
Ravi Shankar asked: I'm stuck on the confusion matrix exercise — how do I know which class should be treated as "positive"?
Message history
Replies you've sent, pending review or already delivered
Sep 9 — Machine Learning
Ravi Shankar asked: Could you go over why we used L2 regularization instead of L1 in the example from class?
Your reply
Great question — L2 shrinks weights smoothly and keeps all features, while L1 can zero some out entirely (useful for feature selection). In our example we wanted to keep every feature but reduce overfitting, so L2 was the better fit.