Hi, According to festival usermanual, poslex contains reverse probabilities of the words with special tag. i make my own poslex but my reverse probabilities are so far from OALD poslex. my questions are: 1. why probabilites are negetive? 2. why they are greater than one? 3. Although the word's frequencies are very different, but the probability range are very close. for example: ("in" ((fw -4.723) (in -1.265) (nnp -10.016) (rb -6.408) (rbr -7.419) (rp -2.830) ) () ) ("inaccurately" ((rb -9.201) ) () ) Have you used any probability normalization approach? Could you please help me, because i've confused completely.
Best Regards, Ahmad
_______________________________________________ Festlang-talk mailing list Festlang-talk@lists.berlios.de https://lists.berlios.de/mailman/listinfo/festlang-talk