A Comparative Study of LSTM and GRU Models for Cryptocurrency Price Forecasting

The price of cryptocurrencies is hard to anticipate because of how dynamic and turbulent the market is. The effectiveness of two popular machine learning models,Gate Recurrent unit (GRU) and Long Short Term Memory (LSTM), was compared in attempt to overcome this problem. Due to their capacities of modeling temporal dependency in sequential data, these models are particularly well suited for estimating price changes based on historical data. This study relied on historical price data concerning cryptocurrencies, so the closing prices for a variety of cryptocurrencies were collected as input data. This is all obtained using the yfinance library which pulls data from Yahoo Finance and gives users options regarding how to obtain financial information from these services. This study relied on closing prices to increase model accuracy and ideally simplify forecasting.