Question

I am trying to extract Arabic proper names from a text using Stanford Parser.

for example if I have an input sentence:

تكريم سعد الدين الشاذلى

using the Arabic Stanford parser, the tree diagram will be:

(ROOT (NP (NN تكريم) (NP (NNP سعد) (DTNNP الدين) (NNP الشاذلى))))

I want to extract the proper name:

سعد الدين الشاذلى

which have the sub-tree:

(NP (NNP سعد) (DTNNP الدين) (NNP الشاذلى))

I have tried this: similar question

but there is some thing wrong in this line:

List<TaggedWord> taggedWords = (Tree) lp.apply(str);

the error in putting a tree type in a list of taggedword
another thing that I didnot understand that where could i use the suggested taggedYield() function

Any Ideas, please?

Answer 1

This is pretty basic Java with respect to the library, but what you want is:

Tree tree = lp.apply(str);
List<TaggedWord> taggedWords = tree.taggedYield();
for (TaggedWord tw : taggedWords) {
  if (tw.tag().contains("NNP")) {
    System.err.println(tw.word());
  }
}