Today: Sunday, 9 March 2025
-
Information Technology
-
Computer
-
Telecommunications
Designing a Novel Hybrid Algorithm Utilizing Dual LSTM & Dual RNN for Financial Market Prediction (Bitcoin)
Volume 5, Issue 1, 2023-2024, Pages 98 - 106
1 Intelligent Systems Laboratory, Computer Engineering Faculty, Amirkabir University of Technology, Tehran
2 Intelligent Systems Laboratory, Computer Engineering Faculty, Amirkabir University of Technology, Tehran
Abstract :
Background: Financial markets, known for their complexity and volatility, require advanced forecasting methods. These predictions are crucial for stakeholders to anticipate significant shifts, such as asset price changes. Experts use a variety of machine learning tools, including quantitative finance and neural networks, to generate reliable market forecasts. Aim: This study focuses on predicting Bitcoin prices using deep learning algorithms. The goal is not perfection, but to enhance predictive accuracy using established methods. Methods: We gather detailed market data via the Binance API, preprocess it, and extract key features. Our model combines LSTM and RNN architectures for Bitcoin price prediction, emphasizing practicality over a flawless system. Results: Using Python and TensorFlow, we test our model, which integrates LSTM and RNN. It outperforms other variants in metrics like mean squared error and mean absolute error, demonstrating its effectiveness in Bitcoin forecasting. We acknowledge its strengths and areas for improvement. Conclusion: Machine learning, especially deep learning, is valuable in financial market analysis, helping to identify complex patterns for asset prediction. While acknowledging its potential, we also recognize its limitations and advocate for open-source collaboration for further advancements.
Background: Financial markets, known for their complexity and volatility, require advanced forecasting methods. These predictions are crucial for stakeholders to anticipate significant shifts, such as asset price changes. Experts use a variety of machine learning tools, including quantitative finance and neural networks, to generate reliable market forecasts. Aim: This study focuses on predicting Bitcoin prices using deep learning algorithms. The goal is not perfection, but to enhance predictive accuracy using established methods. Methods: We gather detailed market data via the Binance API, preprocess it, and extract key features. Our model combines LSTM and RNN architectures for Bitcoin price prediction, emphasizing practicality over a flawless system. Results: Using Python and TensorFlow, we test our model, which integrates LSTM and RNN. It outperforms other variants in metrics like mean squared error and mean absolute error, demonstrating its effectiveness in Bitcoin forecasting. We acknowledge its strengths and areas for improvement. Conclusion: Machine learning, especially deep learning, is valuable in financial market analysis, helping to identify complex patterns for asset prediction. While acknowledging its potential, we also recognize its limitations and advocate for open-source collaboration for further advancements.
Keywords :
Financial Market Prediction, Deep Learning Algorithms, LSTM, RNN
Financial Market Prediction, Deep Learning Algorithms, LSTM, RNN