Conditional entropy
Front
Active users
4
All-time users
4
Favorites
0
Last updated
3 years ago
Date created
Dec 31, 2021
Chapter 3
(2 cards)
Conditional entropy
Quantifies average uncertainty about \(X\) after observing \(Y\):
$$\text{ENT}(X|Y) = \sum_y Pr(y) \text{ENT}(X|y),$$
where
$$\text{ENT}(X|y) = -\sum_x Pr(x|y) \log_2 Pr(x|y).$$
Entropy
Uncertainty about a variable \(X\) is quantified using entropy:
$$\text{ENT}(X) = -\sum_x Pr(x) \log_2 Pr(x),$$
where \(0 \log 0 = 0\) by convention