Figure - available from: PLOS One
This content is subject to copyright.
Source publication
Forecasting speculative stock prices is essential for effective investment risk management and requires innovative algorithms. However, the speculative nature, volatility, and complex sequential dependencies within financial markets present inherent challenges that necessitate advanced techniques. In this regard, a novel framework, ACB-XDE (Attenti...
Citations
... Good Practices for Financial Time Series AI [17] identifies best practices for implementing explainability in AI-driven financial forecasting systems, emphasizing the importance of data quality and tailored methods for specific audiences while considering data properties. CAB-XDE Framework [38] details an innovative decision ensemble framework that combines customized attention BiLSTM with XGBoost for predicting speculative stock prices, validated through empirical analysis in the volatile Bitcoin market and showing superior performance compared to existing models. ...
The growing complexity of price forecasting necessitates advanced methods to improve accuracy. In this study, we introduce a novel approach that integrates multiple deep learning architectures, specifically recurrent neural networks (RNNs), convolutional neural networks (CNNs), and transformers. This integration enables the model to effectively capture both temporal and spatial dependencies in price data. By employing an ensemble strategy, we combine the predictions from these diverse models, thereby reducing biases and variances inherent in individual predictors. Additionally, transfer learning techniques are utilized by incorporating pre-trained models, enhancing generalization on our specific task. Rigorous experiments on various datasets illustrate the effectiveness of our approach, revealing significant advancements in forecasting accuracy over conventional single-model techniques. Our integrated framework serves as a powerful solution to challenges faced in price prediction, facilitating improved decision-making across a range of industries. The evaluations underscore the benefits of deep learning model integration, demonstrating its suitability for practical applications in price forecasting.
