LSTM and GRU deep learning models applied to analyze the effect of scraped past 30-day Coindesk news sentiment on future 10-day volatility for five different cryptocurrencies: Bitcoin, Binance Coin, Ether, and XRP. These models are used together with accumulated local effects to visualize the effects of the sentiment.
The research paper is available here. The code in this repo is mostly based on Jupyter Notebooks and might be incomplete.
