Cryptocurrency Price Prediction Using Deep Learning

Abstract

After the fall in the price of cryptocurrencies in recent years, bitcoin was increasingly considered an investment vehicle. Due to its very unstable nature and frequent changes, there is an increasing need for good predictions of the values on which later investment decisions will be based. In this project, the use of deep learning algorithms is considered to predict the price of bitcoin, taking into account various parameters that affect the value of bitcoin. The proposed system relies on the LSTM architecture to predict the value for the next N days, based on the previous days' value. For the first phase of the research, historical data on cryptocurrency values was used as a knowledge source. In continuation, a trained model receives values for N days backward to predict the value for the desired day. The second phase of the research involves creating a frontend page of the application that allows users to get a prediction of the value of any cryptocurrency they want, for the period they specify.