Kurniasari, Dian and Rizkiansyah, Farrel and Notiragayu, Notiragayu and Warsono, Warsono (2025) Hybrid ARIMA-ANN Model for Cryptocurrency Price Forecasting. International Journal of Multidisciplinary Research and Publications, 7 (7). pp. 131-130. ISSN 2581-6187

Text Published - IJMRAP-V7N7P62Y24.pdf Download (767kB)

Abstract

Investments in digital currencies or other financial instruments during the COVID-19 pandemic were primarily driven by the pursuit of profits. Price forecasting for future periods using time series analysis is a key strategy in achieving this goal. ARIMA, a widely used method in time series analysis, is effective for modeling linear patterns but struggles with non-linear components often present in time series data. To address this limitation, this study introduces a Hybrid ARIMA-ANN approach, designed to capture both linear and non-linear patterns effectively. The proposed method involves two main models and evaluates performance under three data-splitting schemes: 60% training and 40% testing, 70% training and 30% testing, and 80% training and 20% testing. The ARIMA model is applied to forecast the primary data, while the ANN model predicts ARIMA residuals to refine overall accuracy. The optimal results were achieved using the 80%-20% data split, with the ANN model yielding an accuracy of 99.89%, an MSE of 0.0625, an RMSE of 0.2501, and a MAPE of 0.1032. These metrics indicate a highly precise model capable of reliable cryptocurrency price prediction. Based on this model, cryptocurrency prices are forecasted to decline from February 2022 to April 2023. This study demonstrates the potential of Hybrid ARIMA-ANN as a robust forecasting tool, particularly for applications where non-linear data dynamics are prominent. The findings contribute to advancing predictive analytics in financial markets.

Item Type:	Article
Subjects:	Q Science > QA Mathematics
Divisions:	Fakultas Matematika dan Ilmu Pengetahuan Alam (FMIPA) > Prodi Matematika
Depositing User:	DIAN KURNIASARI
Date Deposited:	24 Feb 2025 07:31
Last Modified:	24 Feb 2025 07:31
URI:	http://repository.lppm.unila.ac.id/id/eprint/54435

Actions (login required)

View Item