I have performed two analyses of this dataset.
The first analysis ("Movie_Reviews_Predictions") was my first NLP projects. For it, I built two different models, one using Bag of Centroids with Word2Vec, and the other using TfidfVectorizer.
The second analysis ("RNN_movie_reviews") uses a recurrent neural network and TensorBoard to measure the performance of the various iterations of my model. Here's a link to an article that I wrote about this method: https://medium.com/@Currie32/predicting-movie-review-sentiment-with-tensorflow-and-tensorboard-53bf16af0acf
The data for this project is from a 2015 Kaggle competition: https://www.kaggle.com/c/word2vec-nlp-tutorial. Despite the two years having past, it is still an excellent learning tool. Having the ability to compare my results to others provided a good benchmark and motivation to improve my code. There was also a tutorial, provided by Google, which helped me to create my first analysis.
To view my work most easily, click on the .ipynb files.