Sentiment Analysis Convolutional Neural Network

• Code is run on CPU;

• Implemented in Jupyter Notebook (Python 3.6);

• NLP Classification problem;

• 1D ConvNet model;

• Keras Functional API;

• Code is based on the fast.ai MOOC, but used IMDb Keras dataset directly and has own prediction function to print out results:

  • GloVe pretrained word embeddings;
  • Accuracy (on test set): 0.8937.