Using N-Gram Multichannel CNN for Sentiment Analysis
The model uses multiple parallel convolutional neural networks that read the source document using different kernel sizes.This,in effect,creates a multichannel convolutional network for the text that reads text with different n-gram sizes.
Link to the paper- http://www.people.fas.harvard.edu/~yoonkim/data/sent-cnn.pdf
Used the implementation of models in this paper for Sentiment Analysis on movie reviews dataset.
Link to the dataset - https://www.cs.cornell.edu/people/pabo/movie-review-data/review_polarity.tar.gz