search

quanteda / quanteda.textmodels

Text scaling and classification models for quanteda
42 stars 6 forks source link

Implement native svmlin function #14

Closed kbenoit closed 3 years ago

kbenoit commented 4 years ago

This would eliminate the requirement for the large RSSL package that contains the function we call to implement the Sindhwani and Keerthi (2006)^[1] method for fast linear transductive SVMs. We just need to reproduce https://github.com/jkrijthe/RSSL/blob/master/R/svmlin.R including its C++ code and adapt it for the Matrix classes.

[1] Vikas Sindhwani and S. Sathiya Keerthi. Large Scale Semi-supervised Linear SVMs. Proceedings of ACM SIGIR, 2006 @references V. Sindhwani and S. Sathiya Keerthi. Newton Methods for Fast Solution of Semi-supervised Linear SVMs. Book Chapter in Large Scale Kernel Machines, MIT Press, 2006

koheiw commented 4 years ago

Please do not forget that there is the test-svmlin branch for this.