/SemEval2017Task5

Approach for SemEval-2017 Task 5

Primary LanguagePython

SemEval-2017 Task 5 Fine-Grained Sentiment Analysis on Financial Microblogs and News

http://alt.qcri.org/semeval2017/task5/

Usage example:

-- Cross validation

python main.py --train_file Microblog_Trainingdata.json --mode cv --subtask 1

python main.py --train_file Headline_Trainingdata.json --mode cv --subtask 2

-- Prediction

python main.py --train_file Microblog_Trainingdata.json --test_file Microblogs_Testdata.json --mode predict --subtask 1 --regressor svm

python main.py --train_file Headline_Trainingdata.json --test_file Headlines_Testdata.json --mode predict --subtask 2 --regressor rnn

Train and test data files can be downloaded from SemEval-2017 site