The goal of the project is to predict the sentiment of phrases using the labeled dataset. The data "Rotten Tomatoes movie review dataset" is a corpus of movie reviews used for sentiment analysis. This Kaggle competition has been inspired by Socher's work on labeling parsed phrases in the corpus.
The initial phase of the project will be focusing on getting the best accuracy results using the CNN-LSTM architecture, Then other architectures will be explored to further improve on CNN-LSTM.