SVM Tagger

How to train

• prepare the training file
  • follow instructions for English or French
  • run training file through utf8-html
  • replace PONFP by period if necessary
  • not necessary to delimit sentences by blank lines
• make a config-#.svmt file
• a good file to use is ../MIDDLE-FRENCH/SVMT-config/config-3.svmt
• SVMTlearn {config-#.svmt}

How to tag

• prepare the input file
  • follow instructions for English or French
  • run file through utf8-html
• in the following command, # should match the number from the relevant config file
• SVMTagger -V 2 -S LRL -T 4 cat# < {input} > {output}