A typical IE/NLP pipeline involves sentence segmentation, tokenization, POS tagging, chunking, entity detection and relation extraction tasks. Compare and contrast the functionalities provided following frameworks/toolkit for implementing such a pipeline using the literature:
(a) NLTK, (b) Stanford CoreNLP, (c) Apache OpenNLP (d) SpaCy and (e) GATE
Consider one such task (e.g. POS tagging, chunking or Named entity recognition) and evaluate the performance of two of the above tool kits on a relevant CoNLL data set.
(CoNLL--‐2000 for chunking, CoNLL--‐2002 for NER)
I don't think a perfect article exists. I would start with the following paper and wait for further comparisons (future publications which cite it) or start searching for reports comparing particular features. That would be a nice paper if you survey them all!
http://drops.dagstuhl.de/opus/volltexte/2016/6008/pdf/OASIcs-SLATE-2016-3.pdf
Thank you very much Rafal Rzepka. I need main comparison on NLTK vs Spacy. I you have such article please forward it to me. Any way Thank you very much for your cooperation.
- Badges
- Science topic
- Similar topics
- Publishing