Lab Assignment 9 - Achu0015/BigDataAna-App-SP2K17 GitHub Wiki

                                          # Lab Assignment 9

Solution 1:

Implementing a CNN for Text Classification using Tensorflow. Here the dataset used is Movie Review Dataset from Rotten Tomatoes the link can be found here. This model achieves classification for a range of text. This dataset contains 10k reviews which comprises of half negative and half poitive, and contains around 20k vocabulary.

The first layers embeds words into low-dimensional vectors, then next layer performs convolutions over the embedded word vectors using multiple filter sizes. For example, sliding over 3, 4 or 5 words at a time. Next, we max-pool the result of the convolutional layer into a long feature vector, add dropout regularization, and classify the result using a softmax layer.

After building the model below are the results for accuracy at each 100 iterative checkpoint.

So the accuracy increases with the number of iterations performed.

The Tensorboard visualizations can be seen in below screen shots, for Accuracy and loss over numerous iterations:

Solution 2:

Tried to implement the conversation app but able to do only upto making conversation as I did in lab 5 solution 2. I was unable to deploy in Heroku, I'm getting the error of nothing found when I click the deployment URL.