Convolutional Ensemble Learning for Stock Price Direction Forecasting and Algorithmic Trading
Küçük Resim Yok
Tarih
2025
Yazarlar
Dergi Başlığı
Dergi ISSN
Cilt Başlığı
Yayıncı
Institute of Electrical and Electronics Engineers Inc.
Erişim Hakkı
info:eu-repo/semantics/closedAccess
Özet
Financial time series forecasting is challenging due to the noisy, non-stationary, and volatile nature of market data. In this context, predicting short-term price direction is a critical task for algorithmic trading. While deep learning (DL) models have shown promising results, they often suffer from reproducibility issues, limiting their reliability in real-world financial applications. This study proposes a DL-based framework that predicts the short-term direction of stock price movements using image representations of time series data. To address the instability of individual models, an ensemble learning strategy is employed. Specifically, a VGG16 convolutional neural network was fine-tuned 10 times independently with different random weight initializations, and the final prediction was generated by majority voting across model outputs. Results demonstrate that the proposed framework improves reproducibility and predictive reliability while delivering competitive financial performance when integrated into trading strategies. © 2025 IEEE.
Açıklama
9th International Artificial Intelligence and Data Processing Symposium, IDAP 2025 -- 6 September 2025 through 7 September 2025 -- Malatya -- 215321
Anahtar Kelimeler
CNNs, ensemble learning, financial time series forecasting, transfer learning
Kaynak
9th International Artificial Intelligence and Data Processing Symposium, IDAP 2025
WoS Q Değeri
Scopus Q Değeri
N/A
