BitForecast is a Bitcoin forecasting application that uses the Liquid Time-stochasticity network described here: https://github.com/Ammar-Raneez/Liquid-Time-stochasticity-networks for its forecasting model. Additionally, it considers multiple exogenous factors, including Twitter volume, Google Trends, Twitter sentiment, and the block reward size, alongside the basic historical prices, to produce a more robust and effective forecast.
- Node 16+
- Python 3.8+
- client - frontend application built using React & Redux
- server - Flask API server for local trials & testing
- ml - Machine learning experiments, trails & testing
- deployment - Flask API server hosted in Heroku