Deep Learning Based Forecasting of Cryptocurrency Markets By Using Bitcoin and Ethereum as Case Study for 2016&ndash;2024

Hussein Abd Al-Aali, Mustafa; Kalaf Aziz, Hussein; Kareem Abd Al-Razzaq, Kawthar

doi:10.33916/qjae.2026.01224240

	Deep Learning Based Forecasting of Cryptocurrency Markets By Using Bitcoin and Ethereum as Case Study for 2016–2024
AL-Qadisiyah Journal For Administrative and Economic sciences
Volume 28, Issue 1, April 2026, Pages 224-240 PDF (933.42 K)
Document Type: Research Paper
DOI: 10.33916/qjae.2026.01224240
Authors
Mustafa Hussein Abd Al-Aali; Hussein Kalaf Aziz; Kawthar Kareem Abd Al-Razzaq
University of Wasit
Abstract
The research aims to develop an accurate model for the prediction of Bitcoin and Ethereum values using deep learning methods. These are significant digital assets in today's financial markets. Daily data collected from 2016 to 2023 was analyzed by LSTM and CNN and the combined LSTM-CNN model in order to assess their performance in handling complex time patterns and price shifts, usually characterizing cryptocurrency markets. The statistical results revealed that the LSTM model performed best according to performance metrics such as RMSE and MAE and MAPE and R². Moreover, strong generalization with precision in predictions while avoiding overfitting was depicted. These findings give enhanced statistical and economic performance that involves enhanced prediction accuracy and therefore helps investment and risk management in an evolving financial scenario. It highlights the usage of AI in analyzing digital markets, giving better insight than previous studies dealing either with traditional models or less integrated models. The findings mark progress toward more reliable economic models capable of handling complex dynamic financial data.
Keywords
Cryptocurrencies; Deep Learning; LSTM Models; Bitcoin; and Ethereum

Statistics Article View: 29 PDF Download: 38