Comparison of Prophet and LSTM Model Performance in Forecasting Ethereum Prices Using a Machine Learning Approach
Fintech Innovation Journal · 01.05.2026 · 2 (2) · 136
DOI: 10.63913/ftij.v2i2.28 ↗Kurz gefasst
Fachartikel in der Zeitschrift Fintech Innovation Journal, online erschienen am 01.05.2026. Verfasst von Siva Subramanian R.
Abstract (Originalsprache)
The price of cryptocurrency assets such as Ethereum is highly volatile, which creates challenges in producing accurate forecasts. This study compares the performance of two machine learning models, Prophet and Long Short-Term Memory (LSTM), in predicting Ethereum prices based on daily historical data. Prophet is used to capture long-term trends and seasonal patterns, while LSTM focuses on modeling nonlinear relationships and price fluctuations. The results show that Prophet provides stable trend estimates with narrow confidence intervals but tends to lag during periods of high volatility. In contrast, LSTM performs better in predicting short- to medium-term price dynamics, including sharp increases. Therefore, Prophet is more suitable for long-term trend analysis, whereas LSTM is more effective for modeling nonlinear patterns in the dynamic cryptocurrency market.
Zitieren
APA
R, S. S. (2026). Comparison of Prophet and LSTM Model Performance in Forecasting Ethereum Prices Using a Machine Learning Approach. Fintech Innovation Journal, 2(2), 136. https://doi.org/10.63913/ftij.v2i2.28
Vancouver
R SS. Comparison of Prophet and LSTM Model Performance in Forecasting Ethereum Prices Using a Machine Learning Approach. Fintech Innovation Journal. 2026;2(2):136. doi:10.63913/ftij.v2i2.28
BibTeX
@article{R2026,
author = {R, Siva Subramanian},
title = {Comparison of Prophet and LSTM Model Performance in Forecasting Ethereum Prices Using a Machine Learning Approach},
journal = {Fintech Innovation Journal},
year = {2026},
volume = {2},
number = {2},
pages = {136},
doi = {10.63913/ftij.v2i2.28}
}
EUVIMED