Apply an RNN to the task of text classification, in particular binary sentiment analysis: classifying movie reviews on IMDB as either positive or negative. Run a notebook such as www.tensorflow.org/text/tutorials/text_classification_rnn and report on your results. 

a. What level of accuracy can you achieve? 

b. Gather some reviews from another site, not IMDB, and see if the trained model performs as well on them. 

c. How does accuracy vary with the length of the review? 

d. Which words are most associated with a positive review? A negative review?