Predicting Price of Cryptocurrency and Blockchain in an Uncertain Environment

Abstract

Fueled by innovation and decentralization, the cryptocurrency market presents a captivating yet volatile investment landscape. Accurately predicting cryptocurrency prices in such an environment remains a significant challenge. This paper explores the potential of a hybrid framework combining machine learning and regression analysis to enhance price prediction accuracy while acknowledging the market's inherent uncertainty. We delve into the existing literature on cryptocurrency price prediction, focusing on regression techniques and their limitations. The paper then proposes a novel framework that integrates a Long Short-Term Memory (LSTM) network with a Multiple Linear Regression (MLR) model to capture historical trends and the impact of key market factors. The methodology section details data collection, pre-processing, and model development. Numerical results from a case study involving prominent cryptocurrencies and a comparative analysis of the proposed framework against established methods are presented. Finally, the conclusion discusses the findings, limitations, and future research directions in predicting cryptocurrency prices within the evolving blockchain ecosystem.