sergey-tihon / Stanford.NLP.NET

Stanford NLP for .NET
http://sergey-tihon.github.io/Stanford.NLP.NET/
MIT License
598 stars 123 forks source link

new ideas for c# samples #34

Closed JimSEOW closed 2 years ago

JimSEOW commented 9 years ago

https://medium.com/@klintcho/training-a-swedish-pos-tagger-for-stanford-corenlp-546e954a8ee7

sergey-tihon commented 9 years ago

@Davtomas Do you want to create PR with C# sample for this?)

JimSEOW commented 9 years ago

The stanfordnlp.NET you started is a good platform for others to share http://stackoverflow.com/questions/15609324/training-n-gram-ner-with-stanford-nlp

sdg002 commented 7 years ago

Very useful. Thank you for doing this work.

sdg002 commented 7 years ago

Subject: POS results of C# different from the online version

Hi All, I tried parsing the following sentence using the online page hosted here(http://nlp.stanford.edu:8080/parser/index.jsp). and compared the with the results I get from my custom application.

Test string: The Vendor must not resell,sell, rent, lease, loan, supply, distribute, redistribute, publish or re-publish the Software.

Results form online: not/RB resell/VB ,/, sell/VB ,/, rent/VB ,/, lease/VB ,/, loan/NN ,/, supply/VB

Results from custom C# app using "wsj-0-18-bidirectional-nodistsim.tagger": not/RB resell/VB ,/, sell/VB ,/, rent/NN ,/, lease/NN ,/, loan/NN ,/, supply/NN

Notice that in the latter, the words rent, lease and loan have been classified as a noun. The online version has correctly classified as verb. Please advice if I have missed something.

Thanks, Sdg

sergey-tihon commented 7 years ago

@sdg002 Compare versions

Latest version on NuGet is 3.7.0 published at 2016-10-31 Site still use

Parser: englishPCFG.ser.gz 
Last updated 2016-09-12

It may be difference in versions or in loaded parser (models)

sergey-tihon commented 7 years ago

Results from http://nlp.stanford.edu:8080/corenlp/process screen shot 2017-02-16 at 2 58 18 pm

Results from http://corenlp.run screen shot 2017-02-16 at 2 59 00 pm

sdg002 commented 7 years ago

thanks for looking into this so quickly.

sergey-tihon commented 2 years ago

Close as an old issue