BinHeRunning / disease_name_recognition_through_rnn

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

disease_name_recognition_through_rnn

This directory contain code and data for "Recurrent neural network models for disease name recognition using domain invariant features". ACL-2016, Berlin, Germany, Aug-2016.

Dataset used here is NCBI disease corpus which is publicly available at "http://www.ncbi.nlm.nih.gov/CBBresearch/Dogan/DISEASE/".

binary_data: All input file for disease mention recognition

class_data: All input file for disease mention recogntion with class

word2vec: All word vectors we learned throug CBOW, Skip-gram and GloVe in pubmed corpus. Here each line contain vectors for correspondin word in vocab file.

results and multi_results: results obtain by disease mention and disease class recognition respectivly

Software Requirement: Theano Numpy Cuda 6 (Optional)

PS: If you have GPU in machine it will run automatically in GPU without changing anything in code.

About


Languages

Language:Python 59.3%Language:Perl 40.7%