nekorush14 / gr-rnn-sentence-classification

2017 my graduation research at Hosei Univ, Japan

gr-rnn-sentence-classification

Description

This project is targeted for sentence classification by RNN structure.

Requirements

This project require following environments

Environment	Version
Python	3.6.x or later
TensorFlow	1.4.x - 1.12.0
numpy	1.13.x

Dataset

This project use following dataset for train and test.

• Sentence polarity dataset v1.0 Movie Review data from Rotten Tomatoes

Reference of Dataset

Seeing stars: Exploiting class relationships for sentiment categorization with respect to rating scales

Usage

Learning Step:
Use sentece_classification.py. Once executed, learning is executed with the default hyperparameter and model data and vocabulary are generated under the "run" directory.

Classification App:
Use Classifier.py. Once execute, classification is executed with the default hyperparameter and the classification result of the test sentence written in the program is output.

If you want to print all hyper parameter and runtime parameter, please execute the following command.

python sentence_classification.py --help

or

python Classifier.py --help

License

This software is released under the Apache License 2.0, see LICENSE.