- Do Part C 1 (46.64 KiB) Viewed 40 times
do part c
-
answerhappygod
- Site Admin
- Posts: 899603
- Joined: Mon Aug 02, 2021 8:13 am
do part c
do part c
Classifying entities into pre-defined labels 7355 c. Suppose that an NLP Engine wants to tag the sequence, "change is good" using 3 possible tag A, B and C. The engine has the following probabilities information from training data: P(change(A)=1/5, P(change B)=1/5, P(change|C)=4/5 opos2/07/10 P(is|A)=2/5, P(is|B)=0, P(is|C)=0, P(good|A)=1/10, P(good|B)=1/3, P(good|C)=1/3 P(AJA)=1/5,P(BIA)=1/5,P(CIA)=3/5 P(A/B)=1/5,P(BIB)-8/10,P(CIB)=0 P(AIC)-2/3,P(BIC)=1/3,P(CIC)=0 Assume that all the tags have the same probabilities at the beginning of the sentence (and that is 1/3 each). Find out the best tag sequence using the Viterbi algorithm along with value at each vertex. [5marks] d. Given the emission probabilities and transition probabilities find the correct pos tag for the sentence using HMM. Possible Tags are (VB, TO,
Classifying entities into pre-defined labels 7355 c. Suppose that an NLP Engine wants to tag the sequence, "change is good" using 3 possible tag A, B and C. The engine has the following probabilities information from training data: P(change(A)=1/5, P(change B)=1/5, P(change|C)=4/5 opos2/07/10 P(is|A)=2/5, P(is|B)=0, P(is|C)=0, P(good|A)=1/10, P(good|B)=1/3, P(good|C)=1/3 P(AJA)=1/5,P(BIA)=1/5,P(CIA)=3/5 P(A/B)=1/5,P(BIB)-8/10,P(CIB)=0 P(AIC)-2/3,P(BIC)=1/3,P(CIC)=0 Assume that all the tags have the same probabilities at the beginning of the sentence (and that is 1/3 each). Find out the best tag sequence using the Viterbi algorithm along with value at each vertex. [5marks] d. Given the emission probabilities and transition probabilities find the correct pos tag for the sentence using HMM. Possible Tags are (VB, TO,