|
SPRAAK
|
All Data Structures Namespaces Files Functions Variables Typedefs Enumerations Enumerator Groups Pages
|
SPRAAK
|
Collaboration diagram for SprCwrPerp:Data Fields | |
| int | allocated |
| structure is allocated More... | |
| int | wrd_cnt |
| int | hdr_cnt |
| number of words in the condition More... | |
| int | sent_cnt |
| number of sentences handled More... | |
| int | oov_cnt |
| number of OOV's handled More... | |
| int | cnt |
| number of words used to calculate the perplexity More... | |
| double | pp |
| the perplexity of a sentence/corpus More... | |
| double | ppx |
| the perplexity of a sentence/corpus excluding OOV's More... | |
| SprCwrClassSeqPtr * | opt_seq |
| the best sequence More... | |
| SprCwrLMInterface | lmi |
| interface to the LM More... | |
| SprCwrLMInstance * | lm_data |
| pointer to the LM-data More... | |
| SprCwrSearchLmcHash * | lm_cache |
| check which LM contexts are actively used More... | |
| SprCwrClassSeq * | class_seq_list |
| leaves in the search tree, waiting to be expanded More... | |
| SprCwrClassSeq * | free_list |
| a list of free elements More... | |
| SprCwrClassSeqPtr * | recomb_hash_tbl |
| SprCwrClassSeqPtr * | first_used_slot |
| SprCwrClassSeq * | best_seq |
| int | cnt_ngram [16] |
| int | cnt_steps [16] |
search space, used to find the best word/class sequence for a given LM
| int SprCwrPerp::allocated |
structure is allocated
| int SprCwrPerp::wrd_cnt |
number of words processed in the current sentence (excluding the condition)
| int SprCwrPerp::hdr_cnt |
number of words in the condition
| int SprCwrPerp::sent_cnt |
number of sentences handled
| int SprCwrPerp::oov_cnt |
number of OOV's handled
| int SprCwrPerp::cnt |
number of words used to calculate the perplexity
| double SprCwrPerp::pp |
the perplexity of a sentence/corpus
| double SprCwrPerp::ppx |
the perplexity of a sentence/corpus excluding OOV's
| SprCwrClassSeqPtr* SprCwrPerp::opt_seq |
the best sequence
| SprCwrLMInterface SprCwrPerp::lmi |
interface to the LM
| SprCwrLMInstance* SprCwrPerp::lm_data |
pointer to the LM-data
| SprCwrSearchLmcHash* SprCwrPerp::lm_cache |
check which LM contexts are actively used
only needed if LM contexts must be released
| SprCwrClassSeq* SprCwrPerp::class_seq_list |
leaves in the search tree, waiting to be expanded
| SprCwrClassSeq* SprCwrPerp::free_list |
a list of free elements
| SprCwrClassSeqPtr* SprCwrPerp::recomb_hash_tbl |
a hash table containg the newly expanded sequences, the array is made of tuples (pointer_to_first_hash_el,pointer_to_next_used_slot)
| SprCwrClassSeqPtr* SprCwrPerp::first_used_slot |
pointer to the first used slot (the second element in each slot point to the next used slot)
| SprCwrClassSeq* SprCwrPerp::best_seq |
points to the best sequence (inverse tree) after processing a sentence
| int SprCwrPerp::cnt_ngram[16] |
histogram count on the context length used in prob. calc. on the best path
| int SprCwrPerp::cnt_steps[16] |
histogram count on the nr. of. (typically fall-back) steps taken in prob. calc. on the best path
1.8.6
