yuyichen09 / sentiment-analysis-with-neural-network

This project analyses sentiments using deep learning. I have used only numpy to make the neural network. It portrays basic deep learning features using feed forward, backpropagation, gradient descent and activation functions.

sentiment-analysis-with-neural-network

This project is build up completely with numpy. It implements basic neural network concepts including backpropagation, hidden layers, activation function and gradient descent.

First run 'setup.sh' in bash. This will set up the project for you.

To run the project, open 'run_me.py'. The terminal must show the following:

These are a few screenshots from the project:

Preprocessing

Progress and training

Sentiment analysis

Graphs:

Word Positive/Negative Affinity Distribution

Frequency Distribution Of Words In Our Dataset

Vector T-SNE For Most Polarized Words

Author: Satyaki Sanyal

This project is strictly for educational purposes only.