we should have a didactical chunk that discusses how a super simple linear model can be used for classification, where f = w^T x
so, a simpler model before log. regr.
what we can discuss there is how we can do regression on 0/1 labels, and putting L2 loss / brier score on it.
we should have a didactical chunk that discusses how a super simple linear model can be used for classification, where f = w^T x so, a simpler model before log. regr. what we can discuss there is how we can do regression on 0/1 labels, and putting L2 loss / brier score on it.