Home / Current Issue / Paper 1709196
Development of a Hybrid Deep Learning-Based Model for Bitcoin Prices Prediction
Subject area: Science,Engineering and Technology · Area of research: Deep Learning
Abstract
Bitcoin is a decentralized cryptocurrency, operates on a peer-to-peer network without a central authority. It uses blockchain technology for anonymous and secure transactions. However, the price volatility of Bitcoin is caused by a number of important conditions, such as imbalances in supply and demand attributed to sentiment from shareholders, global events that affect the value of Bitcoin, such as economic recessions or geopolitical disputes, and changes to regulations brought about by organizations and governmental reactions to decentralized currencies. Traditional forecasting techniques struggle to accurately predict Bitcoin's future price due to its high volatility and complexity. This study explores the use of LSTM and GRU neural networks for accurately forecasting Bitcoin prices, aiming to improve investment decision-making in the cryptocurrency market and contribute to the literature on time-series forecasting. The Hybrid LSTM-GRU model performed well with an RMSE of 2268 and MAE of 1693, indicating about 95.17% accuracy in forecasting Bitcoin prices around $35,000. When it comes to Bitcoin price prediction, the Hybrid LSTM-GRU approach outperforms the conventional techniques with greater accuracy and robustness. This model's strong predictive ability has practical applications in finance and cryptocurrency markets, informing investment decisions, market analysis, and research with complex pattern capture for trend prediction
Keywords
Deep Learning, LSTM-GRU model, Bitcoin, cryptocurrency, Prediction.
How to cite this paper
@article{1709196,
author = {Opani M. Aweh, Okebule Toyin, Abiola Kehinde},
title = {Development of a Hybrid Deep Learning-Based Model for Bitcoin Prices Prediction},
journal = {Iconic Research And Engineering Journals},
year = {2025},
volume = {8},
number = {12},
pages = {941-949},
issn = {2456-8880},
url = {https://www.irejournals.com/formatedpaper/1709196.pdf},
abstract = {Bitcoin is a decentralized cryptocurrency, operates on a peer-to-peer network without a central authority. It uses blockchain technology for anonymous and secure transactions. However, the price volatility of Bitcoin is caused by a number of important conditions, such as imbalances in supply and demand attributed to sentiment from shareholders, global events that affect the value of Bitcoin, such as economic recessions or geopolitical disputes, and changes to regulations brought about by organizations and governmental reactions to decentralized currencies. Traditional forecasting techniques struggle to accurately predict Bitcoin's future price due to its high volatility and complexity. This study explores the use of LSTM and GRU neural networks for accurately forecasting Bitcoin prices, aiming to improve investment decision-making in the cryptocurrency market and contribute to the literature on time-series forecasting. The Hybrid LSTM-GRU model performed well with an RMSE of 2268 and MAE of 1693, indicating about 95.17% accuracy in forecasting Bitcoin prices around $35,000. When it comes to Bitcoin price prediction, the Hybrid LSTM-GRU approach outperforms the conventional techniques with greater accuracy and robustness. This model's strong predictive ability has practical applications in finance and cryptocurrency markets, informing investment decisions, market analysis, and research with complex pattern capture for trend prediction},
keywords = {Deep Learning, LSTM-GRU model, Bitcoin, cryptocurrency, Prediction.},
month = {June},
}