- applyWeights
Applies weights to scores produced by language model
- buildUnigramIDMap
Builds the map from unigram to unigramID. Also finds the startWordID and
endWordID.
- clearCache
Clears the various N-gram caches.
- getAvailableBackoff
Selects backoffs for part of word sequence unused in
#getAvailableProb(WordSequence,TrieRange,float)
- getAvailableProb
Selects ngram of highest order available for specified word sequence and
extracts probability for it
- getProbabilityRaw
extracts raw word sequence probability without using caching, making fresh LM
trie traversing
- allocate
- getMaxDepth
Returns the maximum depth of the language model
- getProbability
Gets the ngram probability of the word sequence represented by the word list