Char-RNN, LSTM modeling for Goodreads Reviews

Dates: Jan 2026 - Mar 2026

• Leveraged a Goodreads romance review dataset to create a Character-level Recurrent Neural Network.

• Used machine learning methods from COGS 185 in PyTorch to create the Char-RNN. All coding was in Python, and visualizations were created using Matplotlib.

• Utilized the best models for one- and five-star reviews to generate romance book reviews in the style of Goodreads.

The full report can be found in the Github repository or below.

Report:

    / [pdf]

Project link: https://github.com/ebartt/cogs185final