Skip to content

kyleplathe/Predicting_ETH_Price

Repository files navigation

Group Project 2

presentation_gif

Table of Contents

  1. Hypothesis
  2. Data Collection and Clenaing
  3. Choices for models & Perfromance
  4. Contributions

Hypothesis

  • We use machine learning tools to predict the price of ethereum from historical data, economic indicators, and community sentiment on ethereum specifically from twitter.
  • We test this hypothesis by building LSTM and GRU models

Data Collection & Cleaning

  • Our Sources were:
    • Twint Protocol for Collecting tweet data
    • OWlracle API for collecting Gas price history
    • Kaggle for ETH to USD Historical Data
    • FRED for personal savings percentage data
    • Market Watch for S&P 500 historical data

Choices for models & Performance scores

  • We used two models to predict the price of ethereum

    • Long Short-Term Memory (LSTM) model from keras
    • Gated Recurrent Unit (GRU) model form keras
  • Spliting our data to train and test

GRU Model Peerformance

  • LSTM vs GRU model perfromances
    • LSTM evaluation had 8.98% loss, 91% accuracy LSTM Model Peerformance

    • GRU evaluation had 10.65% loss, 89.35% accuracy GRU Model Peerformance

Contributions

  • Meek Msaki

    • Get historical gas prices and clean data.
    • Set up and run LSTM model
  • Kyle Plathe

    • Get historical eth prices and clean data
    • Get S&P 500 historical
    • Get US savings historical data
    • Set up GRU model
  • Richard Melvin

    • Get historical sentiment for ethereum 2017 - 2020
    • Twitter api sentiment analysis with nltk and vader
    • Perfromed PCA analysis

Presentation & Files

Files:

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published