开发者

Ease of use: Stanford CoreNLP vs. OpenNLP [closed]

开发者 https://www.devze.com 2023-03-18 07:46 出处:网络
Closed. This question is opinion-based. It is not currently accepting answers. Want to improve 开发者_运维技巧this question? Update the question so it can be answered with facts and citation
Closed. This question is opinion-based. It is not currently accepting answers.

Want to improve 开发者_运维技巧this question? Update the question so it can be answered with facts and citations by editing this post.

Closed 9 years ago.

Improve this question

I looking to use a suite of NLP tools for a personal project, and I was wondering whether Stanford's CoreNLP is easier to use or OpenNLP. Or is there another free package you would reccomend? I haven't really done any NLP before, so I am looking for something that I can quickly use to learn the concepts and prototype my ideas. Any help is appreciated.


My opinion on which is easier to use is biased, but regarding Ivan Akcheurov's answer, we only released Stanford CoreNLP in Oct 2010, so it isn't very old. Regarding his suggestions, it seems to depend on whether you want to be using a higher-level processing framework or actual processing tools. E.g., if you poke around Knime, it appears that the only NLP components included are actually OpenNLP ones, and most of the machine learning is wrapping Weka.... For groups of individual tools that work together, Stanford NLP, OpenNLP, NLTK, and Lingpipe are perhaps the main choices.


I suggest you GATE (gate.ac.uk):

GATE

  1. Language: Java
  2. Has UIMA support integrartion
  3. Documentation: Super great documented! Movie tutorials and Training Course
  4. Has GUI
  5. Ability to use WordNet, Lucene, Google, Yahoo, Google Translate, Weka
  6. Has some parts of LingPipe and OpenNLP as a plugin

OpenNLP

  1. Language: Java
  2. SharpNLP (its C-Sharp port)
  3. Has UIMA support integrartion

LingPipe

  1. Language: Java
  2. Documentation: Free book tutorials

NLTK

  1. Language: Python
  2. Documentation: an excellent free book
  3. Corpora: Provides dozen of corpora data (~ 850 MB) and lexicons such as wordnet etc.


I suggest you Stanford as it provides the multiple things under one package that is opensource also e.g. Stanford CoreNLP has

  1. StanFord Parser.
  2. Stanford POS Tagger.
  3. Stanford Named Entity Recognition.
  4. Stanford Typed Dependencies. etc.

So in short under one umbrella you get multiple Solutions....

0

精彩评论

暂无评论...
验证码 换一张
取 消

关注公众号