Convolutional Neural Networks for Sentence Classification

This repository contains the code which was used for the paper Sentiment Analysis of E-Commerce Reviews with Convolutional Neural Network using Natural Language Processing by Rohan Sawant. It started out as fork of Alexander Rakhlin's Convolutional Neural Networks for Sentence Classification in Keras .

Dataset

The dateset used for in this repo could be found on Kaggle.

Code

1. Training the Model

To train a new model set ... in sentiment_cnn.py

model_load = False

grid_search = False

3. Evaluating the Model

To run the model on test set ... in sentiment_cnn.py

model_load = True

grid_search = False

2. Running Grid Search

To run grid search update the grid params in sentiment_cnn.py and set

model_load = False

grid_search = True

also

batch_size = [10]

epochs = [10, 15, 20, 25, 30, 35]

optimizer = ['Adamax']

Conclusion

The model CNN-Adamax-Final was trained with 25 Epochs, 10 Batch Size, Adamax Optimizer and 4 Convolutional Layers. This is where the paper diverged from the material which inspired it. The deeper neural network allowed it to better fit to the data.