[ Skip to the content ]

Institute of Formal and Applied Linguistics Wiki


[ Back to the navigation ]

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
parsery [2007/10/16 19:03]
zeman Unwanted wiki markup.
parsery [2007/10/16 21:55] (current)
zeman N-best parsing with Charniak.
Line 77: Line 77:
 make all make all
 </​code>​ </​code>​
 +
  
  
Line 97: Line 98:
 Both folders, ''​charniak-parser''​ and ''​brown-reranking-parser'',​ have a ''​scripts''​ subfolder with the basic set of ''​parse.pl'',​ ''​cluster-parse.pl'',​ and ''​train.pl''​. These scripts are invoked in much the same fashion as for the Stanford parser (see above). Both folders, ''​charniak-parser''​ and ''​brown-reranking-parser'',​ have a ''​scripts''​ subfolder with the basic set of ''​parse.pl'',​ ''​cluster-parse.pl'',​ and ''​train.pl''​. These scripts are invoked in much the same fashion as for the Stanford parser (see above).
  
-Although ​Charniak parser ​can output n best parses for each sentence, ​the scripts currently do not support that. You would have to go into the scripts ​and add ''​-N5'' ​(replace 5 by the N you want) to the ''​parseIt'' ​command line. The output ​will also show you the log prob of the parse.+The ''​parse.pl''​ and ''​cluster-parse.pl''​ scripts of the Charniak parser ​accept ​the ''​-Nbest''​ option, in addition ​to standard options of these scripts''​-Nbest 50'' ​translates as ''​-N50''​ on Charniak'​s ​''​parseIt'' ​commandline. It asks the parser to output N (here 50) best parses, instead of just one. The output ​format for N>1 differs from the default: the set of parses is preceded by a line with the number of parses and the ID (number) of the sentence, and every parse is preceded by a line with the weight (log probability) ​of the parse. This option only applies to ''​charniak-parser''​. It is ignored by ''​brown-reranking-parser''​.
  
 === Training === === Training ===

[ Back to the navigation ] [ Back to the content ]