Nagaraj Naik
Manipal Institute of Technology, Manipal Academy of Higher Education, Manipal, Karnataka, India.

Trisha Bhagat
Department of Software Systems and Cybersecurity, Monash University, Clayton Campus, Australia.

DOI https://doi.org/10.33889/IJMEMS.2026.11.5.096

  ;

Abstract

Bitcoin has gained greater attention as an investment asset in the past few years because of the tremendous price swings in cryptocurrencies. However, its volatility makes it difficult to make accurate predictions that inform investment decisions in the cryptocurrency market. Many researchers and academics have utilized machine learning and deep learning algorithms for prediction. Therefore, this study proposes ensemble deep learning models that combine Long Short-Term Memory (LSTM), Bidirectional LSTM (BiLSTM), Gated Recurrent Unit (GRU), Convolutional Neural Network (CNN), and transformer architectures. In this study, we considered three ensemble models. The first ensemble deep learning model combined LSTM, GRU, and CNN; the second combined BiLSTM, GRU, and CNN; and the third combined CNN, LSTM, and Transformer models. The objective was to capture sequential dependencies and complex temporal patterns in the Bitcoin price data more effectively. For the experimental setup, we collected data from Yahoo Finance between 2015 and 2024. The Mean Absolute Percentage Error (MAPE) and Weighted Mean Absolute Percentage Error (WMAPE) were used to evaluate the model performance. The results demonstrate that the ensemble BiLSTM, GRU, and CNN models outperformed the others, achieving the lowest error rates with an MAPE of 3.62% and a WMAPE of 3.78%.

Keywords- Bitcoin price prediction, Cryptocurrency forecasting, Ensemble deep learning, Financial technology, Economic forecasting, Investment decision support, Financial market stability.

Citation

Naik, N., & Bhagat, T. (2026). Ensemble Deep Learning Models for Bitcoin Price Prediction Using BiLSTM, GRU and CNN. International Journal of Mathematical, Engineering and Management Sciences, 11(5), 2381-2398. https://doi.org/10.33889/IJMEMS.2026.11.5.096.