Artificial Intelligence and Communication Technologies
A Deep Learning CNN based Approach to the Problem of Crypto-Currency Prediction
Authors: Sajan Kumar Kar
Publishing Date: 04-06-2023
ISBN: 978-81-955020-5-9
Abstract
After the invention of Bitcoin by a man named Satoshi Nakamoto along with other blockchain-based person-to-person payment systems, the cryptocurrency market has instantly gained popularity. Because of this, that is, the volatility of the various cryptocurrency prices. This attracts much attention from both the investors and the researchers. The task of forecasting the prices of crypto-currencies because of the static prices and the arbitrary effects in the market is quite challenging. Cryptocurrency price forecasting models that are available now mainly focus on analyzing extrinsic factors, like macro-financial indicators, data linked to the blockchain, and data from social media – with the goal of enhancing the prediction accuracy. However, the intrinsic noise present in the raw data, caused by market and political conditions worldwide, is complex to interpret. In our research we propose a multiple input convolutional neural network model, specifically a convolutional neural network model for the prediction of future cryptocurrency price. Generally, RNNs and LSTMs are used for problems dealing with timeseries data. We used the concept of residual networks on 1-Dimensional convolutional networks to solve the problem of predicting the price of Bitcoin, the most popular cryptocurrency out there at the moment. Furthermore, we conduct additional experiments on ether, the cryptocurrency of Ethereum to further confirm that even CNNs can work equally well, if not better in comparison to the widely used LSTM neural network models.
Keywords
Deep Learning, Cryptocurrency price prediction, CNN.
Cite as
Sajan Kumar Kar, "A Deep Learning CNN based Approach to the Problem of Crypto-Currency Prediction", In: Saroj Hiranwal and Garima Mathur (eds), Artificial Intelligence and Communication Technologies, SCRS, India, 2023, pp. 1123-1136. https://doi.org/10.52458/978-81-955020-5-9-106