Srikanth Pagadala

Sentiment Analysis from Scratch

05 Feb 2017

In this project, we will build a deep network from scratch that classifies movie reviews as positive or negative just based on their text!

Project Outline:

  • Curate a Dataset
  • Developing a “Predictive Theory”
  • STEP 1: Quick Theory Validation
  • Transforming Text to Numbers
  • STEP 2: Creating the Input/Output Data
  • Putting it all together in a Neural Network
  • STEP 3: Building our Neural Network
  • Understanding Neural Noise
  • STEP 4: Making Learning Faster by Reducing Noise
  • Analyzing Inefficiencies in our Network
  • STEP 5: Making our Network Train and Run Faster
  • Further Noise Reduction
  • STEP 6: Reducing Noise by Strategically Reducing the Vocabulary
  • Analysis: What’s going on in the weights?

Source Code

Report