It has been about a year and a half since my Comparing NLP APIs for Entity Extraction post, which continues to be quite popular. I immediately started writing a followup, and naturally never finished or published it.
Since I’ve been getting questions about it lately, I hereby present my entirely unorganized notes that would have been used to actually do some research and finish the followup:
http://en.wikipedia.org/wiki/Named_entity_recognition
http://developer.zemanta.com/docs/Zemanta_API_companion
http://developer.zemanta.com/docs/suggest_markup/
amplify has some improvements, but you need to slightly change the API call… need to compare it to the older one
calais updated some stuff http://www.opencalais.com/documentation/opencalais-documentation/opencalais-release-notes … do I need to relook?
new features on alchemyapi http://blog.alchemyapi.com/?p=70
other changes
* note evri non-commercial use and that you can’t store longer than 2 weeks
* look into opencalais having ’semantic links’ (read the comments)
new APIs
*Wingify doesn’t really do what I want but may be of interest to others
http://www.wingify.com/contextsense/; API details if you contact them…
*uclassify.com let’s you do some simple classification based on giving it some training data; there are
a number of sites like this but not what I’m looking for
comment on original blog post by Andraz Tori
* no api, but http://lcl2.di.uniroma1.it/termextractor/
http://labs.translated.net/terminology-extraction/
* Apache UIMA incubator project
* http://expert1.expertsystem.us.com/essex
http://www.chailabs.com/platform/ and www.inform.com – two examples with no public APIs
http://blog.viewchange.org/2010/05/entity-extraction-content-api-evaluation/#overview
http://affy.blogspot.com/2009/10/entity-extraction-links.html
Pingback: Comparing NLP APIs for Entity Extraction | Michael Fagan's blog