Persian Dependency Parser for Stanford Dependency Parser from a Persian dependency treebank.

Question

I have downloaded the UPDT Persian treebank (Uppsala Persian Dependency Treebank) and I am trying to build a dependency parser model using MaltParser, but I am very new to NLP field and need some help.

Does anyone know how to use MaltParser to build the dependency parser model? and can the new created model be used with the Stanford Dependency Parser in my Java program?

Thank you

See github.com/alvations/nltk/blob/develop/nltk/parse/malt.py#L240 — alvas
MaltParser models aren't compatible with the Stanford Dependency Parser. You'd need to separately train a model for the Stanford Dependency Parser. But MaltParser is also written in Java, so you should be able to call it in a Java program. — Christopher Manning
Thank you for your explanation, now I can train the model, the result is a file like myModel.mco. I want to use the model in my Java program to parse a sentence, but according to theexamples provided by the MaltParser in this directory(maltparser-1.8\examples\apiexamples\srcex\org\maltparser\examples), the sentence must be tokenized and annotated manually, then it can be parsed by the model. Is it the way that Maltparser model work inside Java? Is it possible to make the sentence as input to the model, without tokenizing or annotating? — Dilman Salih

user3639557 user3639557 · Accepted Answer · 2016-02-08T12:49:23

0

votes

See training a parsing model section of http://www.maltparser.org/userguide.html

Persian Dependency Parser for Stanford Dependency Parser from a Persian dependency treebank.

1 Answers