Predicting Ethereum Prices using a LSTM Recurrent Network

Carolyne Pelletier

Akhil Dalal

Tested on: Windows 10

Using the following versions: tensorflow 1.3.0, keras 2.1.2

INSTALL THESE BEFORE RUNNING THE NETWORK.

Libraries required (using pip install):

• pandas
• bs4
• lxml
• tensorflow
• keras
• scikit-learn
• matplotlib

Command to install the above: pip install pandas bs4 lxml tensorflow keras scikit-learn matplotlib

To run: Python3 lstm.py

This will automatically start the training using our best model as outlined in the report.

The network will generate 2 graphs in the current working directory: one for the training/testing data split, and another for predictions.