2019-04-17 Segev, Michael Jacquier, Pierre Han, Zhenze
Models and experiments are split in seperate python scripts that all use common classes to load files and save models to file.
-
NaiveBayesBench.py Run this file to test different feature extraction pipelines with a NB classifier on Stanford Sentiment Treebank.
-
SVMBench.py Run this file to train and test/validate Support Vector Machine model on Stanford Sentiment Treebank.
-
RNNBench.py Run this file to train and test/validate Recursive Neural Network model on Stanford Sentiment Treebank.
-
DecisionTreesBench.py Run this file to train and test/validate extremely random trees model on Stanford Sentiment Treebank.
-
metaClassifier.py Run this file to train and test/validate stacking ensemble meta-classifier on on Stanford Sentiment Treebank using pre-trained models saved as pickle files.
@inproceedings{le2014distributed,
title={Distributed representations of sentences and documents},
author={Le, Quoc and Mikolov, Tomas},
booktitle={International conference on machine learning},
pages={1188--1196},
year={2014}
}
re
numpy
scikit-learn
keras
tensorflow-gpu