This project aims to use a variety of ML and DL tools to predict the price and/or points awarded to a wine. Specifically, I use:
- Basic feature engineering and linear regression
- Latent Dirichlet Allocation (LDA) for topic extraction
- Word2Vec and Long Short-Term Memory (LSTM)
You can try out the app here!
The dataset can be found on Kaggle. It contains about 130,000 wine reviews published to WineEnthusiast with features such as the country, variety, price, points, etc.