NL Reference Corpus

From UNL Wiki
(Difference between revisions)
Jump to: navigation, search
(Files)
(Methodology)
Line 12: Line 12:
 
#The Base Corpus must be '''segmented''' (in sentences) and '''tagged''' for POS.
 
#The Base Corpus must be '''segmented''' (in sentences) and '''tagged''' for POS.
 
#The segmented corpus is used to calculate the '''average sentence length''' (ASL), which is the median of the length (in words) of all sentences.
 
#The segmented corpus is used to calculate the '''average sentence length''' (ASL), which is the median of the length (in words) of all sentences.
#The tagged corpus is used to extract the [[SSS|syntactic surface structures]] (SSS), which are sequences of POS, and to calculate their frequency of occurrence.
+
#The tagged corpus is used to extract the [[LSS|linear sentence structures]] (LSS), which are sequences of POS, and to calculate their frequency of occurrence.
#The average sentence length (ASL) and the syntactic surface structures (SSS) are used to generate the '''NC templates''', as follows:
+
#The average sentence length (ASL) and the linear sentence structures (LSS) are used to generate the '''NC templates''', as follows:
#*NC-A1 = 500 most frequent SSS's where length < (ASL/2) (500 most frequent shortest syntactic structures)
+
#*NC-A1 = 100 most frequent LSS's for noun phrases where length < (ASL/2) (100 most frequent shortest NP's)
#*NC-A2 = 1,000 most frequent SSS's where length < (ASL/2) (1,000 most frequent shortest syntactic structures)
+
#*NC-A2 = 300 most frequent LSS's for verbal phrases where length < (ASL/2) (300 most frequent shortest VP's)
#*NC-B1 = 2,000 most frequent SSS's where length < ASL (2,000 most frequent short syntactic structures)
+
#*NC-B1 = 500 most frequent LSS's where length < ASL (500 most frequent short structures)
#*NC-B2 = 3,000 most frequent SSS's where length < ASL (3,000 most frequent short syntactic structures)
+
#*NC-B2 = 500 most frequent LSS's where length < ASL (500 most frequent short structures)
#*NC-C1 = 4,000 most frequent SSS's  
+
#*NC-C1 = 500 most frequent LSS's where length >= ASL
#*NC-C2 = 5,000 most frequent SSS's
+
#*NC-C2 = 500 most frequent SSS's where length >= ASL
#The NC templates are used to compile the NC corpora: the training corpora and the testing corpora. The training corpora consists of 1 exemplar of each SSS, and will be used to prepare the grammar. The testing corpora consists of 4 exemplars of each SSS randomly selected in the Base Corpus. The whole NC corpora (i.e., 5 exemplars for each SSS) is used to calculate the [[F-measure]], which is the parameter for assessing the precision and the recall of the grammars.
+
#The NC templates are used to compile the NC corpora: the training corpora and the testing corpora. The training corpora consists of 1 exemplar of each LSS, and will be used to prepare the grammar. The testing corpora consists of 4 exemplars of each LSS randomly selected in the Base Corpus. The whole NC corpora (i.e., 5 exemplars for each LSS) is used to calculate the [[F-measure]], which is the parameter for assessing the precision and the recall of the grammars.
  
 
== Files ==
 
== Files ==

Revision as of 19:28, 3 July 2013

The NL Reference Corpus (NC) is the corpus used to prepare and to assess grammars for sentence-based UNLization. It is divided in 6 different levels according to the Framework of Reference for UNL (FoR-UNL):

  • NC-A1: NL Reference Corpus A1
  • NC-A2: NL Reference Corpus A2
  • NC-B1: NL Reference Corpus B1
  • NC-B2: NL Reference Corpus B2
  • NC-C1: NL Reference Corpus C1
  • NC-C2: NL Reference Corpus C2

Methodology

As a natural language corpus, the NC varies for each language. It is derived from a base corpus to be compiled and processed according to the following criteria:

  1. The Base Corpus must have at least 5,000,000 tokens (strings isolated by blank space and other word boundary markers). It must be representative of the contemporary standard use of the written language, and should include documents from as many different genres and domains as possible.
  2. The Base Corpus must be segmented (in sentences) and tagged for POS.
  3. The segmented corpus is used to calculate the average sentence length (ASL), which is the median of the length (in words) of all sentences.
  4. The tagged corpus is used to extract the linear sentence structures (LSS), which are sequences of POS, and to calculate their frequency of occurrence.
  5. The average sentence length (ASL) and the linear sentence structures (LSS) are used to generate the NC templates, as follows:
    • NC-A1 = 100 most frequent LSS's for noun phrases where length < (ASL/2) (100 most frequent shortest NP's)
    • NC-A2 = 300 most frequent LSS's for verbal phrases where length < (ASL/2) (300 most frequent shortest VP's)
    • NC-B1 = 500 most frequent LSS's where length < ASL (500 most frequent short structures)
    • NC-B2 = 500 most frequent LSS's where length < ASL (500 most frequent short structures)
    • NC-C1 = 500 most frequent LSS's where length >= ASL
    • NC-C2 = 500 most frequent SSS's where length >= ASL
  6. The NC templates are used to compile the NC corpora: the training corpora and the testing corpora. The training corpora consists of 1 exemplar of each LSS, and will be used to prepare the grammar. The testing corpora consists of 4 exemplars of each LSS randomly selected in the Base Corpus. The whole NC corpora (i.e., 5 exemplars for each LSS) is used to calculate the F-measure, which is the parameter for assessing the precision and the recall of the grammars.

Files

Software