Hi!

I prefer NLTK because there is a book with a great set of examples.

Also, many corpora are there with  simple methods for parsing, navigating and analyzing.

Lexical resources also are found, such as wordNet, frameNet, and etc.

Regular expressions and many syntax analyzers.

How about alchemyAPI?

Thanks for sharing your thoughts, Mohamed. I agreed, NLTK is great. I love the book, its flow and snippets of working code. However, when it comes to attempting to complete some projects/analyses that should in fact help the students see the final results and drive a point home, NLTK can be limiting (e.g., only first 10 collocations can be printed out, etc.). Also, at times NLTK can be capricious, esp. for 64 bit computers as it depends on MatLab for visualizations.

AlchemyAPI, an IBM spin off? Interesting. Thanks for the tip, Adam. I'm looking into it. It seems mostly geared towards businesses with extremely large datasets. My sense is that they do extraction, categorization, and perhaps sentiment analysis on social media. Is that roughly right? What's your experience with them, Adam?

There's the OpenNLP and Stanford NLP libraries for Java and the Solr extension for php. There are several examples of these libraries online that you can test. The Stanford Foundations of Statistical Natural Language Processing still has a tutorial online at: http://www-nlp.stanford.edu/fsnlp/.

http://www-nlp.stanford.edu/fsnlp/

Hi Mario,

I think it is not free

Have you tried GATE (https://gate.ac.uk/)?

It's Java based rather than python-based, but in my experience is a good teaching tool.

It's open source, freely available, with a good academic pedigree (from the Univeristy of Sheffield's NLP group) and supportive community of users/developers. Key difference is it comes with a GUI (aka IDE) which is a benefit for non-experienced programmers (or people who aren't comfortable with command line dev tools etc).

Hi Tony,

I thing GATE is for text mining. I think there is no lexical analysis and syntax analysis specially in other natural languages. 

To sum it up, what I'm hearing so far is that I have just 2 free Java-based alternatives to NLTK: the Stanford's OpenNLP with extensions (based on Manning and Schutze's book) and the Sheffield's GATE. An IBM's spin off is for-profit and geared towards "big data". Plus I heard one vote for continued use of the Python-based NLTK.

Thanks so much for your input! If you have any further ideas for teaching tools for NLP, please don't hesitate to leave another note.

Thanks, Niall, good to hear from you. I will look into that as well.

Alchemy API was a standalone product,it was recently aquired by IBM.It is a bit too simple.

Thanks, Olga!

If you need good geoparsing, look at http://www.metacarta.com. I do not know whether they still offer at API, but their geoparsing was the best. They did not only parse text, but also assigned relevance ranking to geographic concepts.  

XEROX also has APIs: https://open.xerox.com/Services/XIPParser.

Dear Sascha,

Can you encourage us to use pattern by naming these few tasks.

Thanks! 

Mohamed: you can use GATE for a lot more than just text mining. I've used it for sentiment analysis, named entity recognition, anaphora resolution, and it has plugins for lots more. Also it has good multilingual capability: the default IE system (i.e. ANNIE) is aimed at English but the framework is fully extensible to multilingual data (see https://gate.ac.uk/projects.html)

Thank you for your advise Tony

Apart from NLTK and Pattern, the only other Python NLP library that I've come across is TextBlob ( http://textblob.readthedocs.org/en/dev/ ):

"TextBlob is a Python (2 and 3) library for processing textual data. It provides a simple API for diving into common natural language processing (NLP) tasks such as part-of-speech tagging, noun phrase extraction, sentiment analysis, classification, translation, and more."

I've not used it extensively, so I can't comment on whether or not it would be a good alternative to pure NLTK (TextBlob is apparently built on top of both NLTK and Pattern). It does seem to attempt to simplify the code for certain aspects of text processing such as part-of-speech tagging, so it may suit your teaching needs as an introductory tool.

Thanks to everyone who has been contributing ideas so far! VR