[ Skip to the content ]

Institute of Formal and Applied Linguistics Wiki


[ Back to the navigation ]

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
courses:rg:maxent-lm [2011/04/13 14:05]
popel typo
courses:rg:maxent-lm [2013/04/27 22:52] (current)
popel typo
Line 10: Line 10:
   * Mějme diskrétní prostor jevů X a dvě pravděpodobností rozdělení <latex>P, Q : X \rightarrow [0,1]</latex>. Pak   * Mějme diskrétní prostor jevů X a dvě pravděpodobností rozdělení <latex>P, Q : X \rightarrow [0,1]</latex>. Pak
     * Entropy <latex>H(P) = - \sum_{x \in X} P(x)\cdot \log_2 P(x)</latex>     * Entropy <latex>H(P) = - \sum_{x \in X} P(x)\cdot \log_2 P(x)</latex>
-    * Perplexity <latex>PPL(P) = 2^H(P)</latex>+    * Perplexity <latex>PPL(P) = 2^{H(P)}</latex>
     * Cross-entropy <latex>H(P, Q) = - \sum_{x \in X} P(x) \cdot \log_2 Q(x)</latex>     * Cross-entropy <latex>H(P, Q) = - \sum_{x \in X} P(x) \cdot \log_2 Q(x)</latex>
     * Kullback-Leibler divergence <latex>D_{KL}(P || Q) = \sum_{x \in X} P(x)\cdot\log_2(\frac{P(x)}{Q(x)})</latex>     * Kullback-Leibler divergence <latex>D_{KL}(P || Q) = \sum_{x \in X} P(x)\cdot\log_2(\frac{P(x)}{Q(x)})</latex>

[ Back to the navigation ] [ Back to the content ]