Reuters-21578 Text multi-class multi-label Classification

Text classification with Reuters-21578 datasets using Keras

Downloaded the from reuters21578 data set first. https://archive.ics.uci.edu/ml/machine-learning-databases/reuters21578-mld/

• Tested with Python 3.5
• Keras with Tensorflow 1.3 backend

Those are the top 20 categories we will use for the classification

['pl_usa', 'to_earn', 'to_acq', 'pl_uk', 'pl_japan', 'pl_canada', 'to_money-fx', 'to_crude', 'to_grain', 'pl_west-germany', 'to_trade', 'to_interest', 'pl_france', 'or_ec', 'pl_brazil', 'to_wheat', 'to_ship', 'pl_australia', 'to_corn', 'pl_china']