Menu

Suppose Given True Data Generating Distribution D Binary Classification Problem True Outpu Q43861822

Suppose you are given the true data generating distribution, D,for a binary classification problem (so the true output y is either+1 or 1).

(a) If your loss is absolute loss (|y f (x)), is the Bayesoptimal classifier still optimal? Explain your reasoning or show acounterexample.

(b) If your loss is squared loss ((y f (x))2), is the Bayesoptimal classifier still optimal? Explain your reasoning or show acounterexample.

Expert Answer


Answer to Suppose you are given the true data generating distribution, D, for a binary classification problem (so the true output …

OR