Cryptocurrency Price Prediction using Deep Learning

Cryptocurrency Price Prediction using Deep Learning

Keerthi Aluvala, Uppu Venkat Teja, Gadiputi Madhava Kalyan, Dheeraj Choudhary, Mohd Abid Ali, Swarupa Murala,

Sharath Kumar Reddy Guvvala, Partha Sai Aurangabadkar

Abstract— Cryptocurrency, a virtual or digital currency that are decentralized networks based on blockchain technology, have become increasingly popular in the recent years. Some of the more famous cryptocurrencies include Bitcoin, Ethereum, YFI Ripple, Dash, Iota, Monero, Stratis, Waves etc. Cryptocurrencies have risen from a gray and black market transaction medium to a more vibrant way of investment, becoming increasingly famous among investors and researches alike. This section of the financial market can be labelled as “high risk high reward”. As the crypto market is seeing a 200% growth every year, soon cryptocurrency trading will become a trend in itself. This report documents how our team predicts prices of cryptocurrencies using a Recurrent Neural Network architecture called Long Short Term Memory (LSTM) Algorithm which works on the principle of backpropagation, avoiding long term dependency problem while also optimizing and enhancing the accuracy of the algorithm.

Index Terms— Cryptocurrency, price prediction, Machine Learning, Deep Learning, LSTM, RNNs