Hello, You need to create a new head rules file for German manually.
If you treat German NPs differently too, you can create a new HeadRules.java for German the way it was built for English changing the constituent tags according to the tagset and then using those HeadRules to train the model. Otherwise take a look at the HeadRules interface and implement a simple version of it to read from the head rules file and implement the getHead() and getPunctuationTags functions. Cheers, Rodrigo On Thu, Jun 13, 2013 at 11:53 AM, Tabea Schröder <[email protected]> wrote: > Thanks for that quick help. Now it works for englisch texts. > > But actually i need it for german texts. > Do you know what i have to change that it works for another > language\german? > > Am 13.06.2013 10:24, schrieb Rodrigo Agerri: > >> Hello, >> >> For which language? >> >> The head rules are based on Michael Collins's thesis. Check Annex A >> for explanation: >> >> >> https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=3&cad=rja&ved=0CD0QFjAC&url=http%3A%2F%2Fwww.dfki.de%2F~neumann%2Fdop-seminar%2FReferences%2Fcollins-thesis.pdf&ei=Zn65Ufm1O8OChQeaw4HgDA&usg=AFQjCNEItDe95bYHgAVCvnBWq8Abmlq_PA&sig2=EnSnLqOK0gSVY-PIoZgFcw&bvm=bv.47883778,d.d2k >> >> English head rules for NP are harcoded in >> opennlp.tools.parser.lang.en.HeadRules.java. In this class also you >> will see that HeadRules is read from a rules file into a >> HashMap<String,HeadRule>. >> >> HTH, >> >> Rodrigo >> >> >> On Thu, Jun 13, 2013 at 9:21 AM, Tabea Schröder >> <[email protected]> wrote: >>> >>> Hallo, >>> >>> im trying to train a Parser with the following Method: >>> / opennlp.tools.parser.chunking.Parser.train/(String <*2> >>> languageCode, >>> ObjectStream <*3><Parse <*4>> parseSamples, HeadRules <*5> rules, >>> TrainingParameters <*6> mlParams) >>> >>> But i don't now how to create the *Headrules *Objekt. >>> I know how the Data head-rules.headrules has to look like, but I can't >>> make >>> an Objekt with this Data. >>> >>> Thanks in advance >>> >>> Tabea > >
