Inspiration
A successful business knows what its customers need and feel about their products. Therefore we developed a tool for generating Emotion Analysis about a business.
What it does
Using Long Short Term Memory (LSTM) we are training a model to classify emotional state of a person writing that review. Using those classifications we are providing analysis report about emotion distribution of that business
How we built it
We used tensorflow library to train LSTM to classify emotional state. We used SemEval 2007 Task 14 for training data. Using Flask we created web application's architecture User review about a business were collected by scraping Yelp website. Reviews were classified using trained LSTM model and report was generated using Matplotlib.
Challenges we ran into
Preprocessing the training data to make the data usable for training purpose Modifying LSTM model parameters for improving accuracy
Accomplishments that we're proud of
Learned Tensorflow and implemented LSTM
What we learned
Tensorflow LSTM RNN
What's next for Emotion Analysis for Yelp reviews
Deploying for different business in different domain. For example: For Amazon Products
Built With
- beautiful-soup
- flask
- matplotlib
- numpy
- python
- tensorflow
Log in or sign up for Devpost to join the conversation.