SavedAdd your journal

Journals / Pamukkale Üniversitesi Mühendislik Bilimleri Dergisi / 2026 / Cilt: 32 Sayı: 6

Characterization of common temporal features for forecasting of financial time series

Finansal zaman serilerinin tahmininde yaygın zamansal özelliklerin karakterizasyonu

Ata Eren Arslan, Selami Beyhan

Pages
1293–1303
Publication Date
2026-10-08
🏛️Bibliographic metadata and full text access for this article are provided via DergiPark.
DergiPark Record ↗

Abstract

Purpose: In forecasting of financial time series, stock prices have similar dynamics due to the sectoral dependencies and common disturbances from the financial market. Therefore, associated stock prices can be used to create features in addition to the past values of the target stock prices. Methods: In this study, a novel common dynamics analysis is proposed between financial time series using five temporal patterns such as seasonality, regime switching, trend, volatility and instantaneous change. These potential dynamics are computed over sliding windows, and a binary covariance matrix is constructed based on the direction of their changes. Subsequently, rule mining is applied to construct the ruleset with highest confidence values where stock prices in rule antecedents are used to construct the feature set. A multiple-input single-output model is proposed and trained using fundamental Long Short-Term Memory (LSTM) architecture. Results: For the unknown future horizons, the values of input stock prices are predicted autoregressively by separate LSTM models since their values are needed when forecasting the primary target stock price values. Conclusion: As a summary, this paper has two major contributions. First, stock prices are analysed with common temporal pattern functions that they provide rule sets for input features. Second, the forecasting performance is highly improved using common dynamics such that the application results confirm that the proposed method improves the forecasting performance up to 28% using the same model.

Özet

Amaç: Finansal zaman serilerinin tahmininde, hisse senedi fiyatları sektörel bağımlılıklar ve finansal piyasalardan kaynaklanan ortak dalgalanmalar nedeniyle benzer dinamikler göstermektedirler. Bundan dolayı, ilişkili hisse senetlerinin fiyatları geçmiş değerlerine ek olarak özellikler üretiminde kullanılabilir. Yöntem: Bu çalışmada, finansal zaman serilerinde bulunan beş zamansal desen (mevsimsellik, rejim değişimi, trend, oynaklık ve ani değişim) kullanılarak yeni bir ortak dinamik analizi önerilmektedir. Bu potansiyel dinamikler kayan pencereler üzerinden hesaplanmakta olup değişim yönlerinin değerlerine bağlı olarak ikili bir kovaryans matrisi oluşturulmaktadır. Ardından en yüksek kural değerine sahip kural kümeleri hakkında bilgi sahibi olmak amacıyla kural madenciliği uygulanmakta, bu kurallardaki öncül hisse senetlerinin fiyatları ise özellik kümesinin oluşturulmasında kullanılmaktadır. Bu doğrultuda, temel bir Uzun Kısa Süreli Bellek (LSTM) kullanılarak çok girişli tek çıkışlı bir model önerilmiş ve uygulanmıştır. Bulgular: Bilinmeyen gelecek görüşleri için, giriş hisse senedi fiyatları, temel hisse senetleri fiyatları tahmininde gerekli olduğundan, ayrı LSTM modelleri ile otoregresif olarak tahmin edilmiştir. Sonuç: Özetle, bu çalışmanın iki ana katkısı vardır. Birincisi, hisse senedi fiyatları, giriş özellikleri için kural kümelerini sağlayan ortak zamansal desen fonksiyonları ile analiz edilmiştir. İkincisi, tahminleme performansı ortak dinamikler ile geliştirilmiştir. Uygulama sonuçları göstermektedir ki önerilen yöntemin aynı model ile kıyasla performansı %28 oranında artırdığını doğrulamaktadır.

Keywords
🔖

Cite This Article

Arslan, A. E., Beyhan, S. (2026). Characterization of common temporal features for forecasting of financial time series. Pamukkale Üniversitesi Mühendislik Bilimleri Dergisi, Cilt: 32 Sayı: 6, 1293–1303. https://doi.org/10.65206/pajes.68246

📄 Full Text PDF Viewer

References (25)

  1. [1] Teräsvirta T, Tjøstheim D, Granger CWJ. “Modelling Nonlinear Economic Time Series”, 1st ed. Oxford, Oxford University Press, 2010.
  2. [2] Agrawal R, Imielinski T, Swami A. “Mining association rules between sets of items in large databases”. ACM SIGMOD International Conference on Management of Data, Washington DC, USA, 25-28 May 1993.
  3. [3] Agrawal R, Srikant R. “Mining sequential patterns”. Proceedings of the 11th International Conference on Data Engineering (ICDE), Taipei, Taiwan, 6-10 March 1995.
  4. [4] Fournier-Viger P, Wu CW, Tseng VS, Cao L, Nkambou R. “Mining partially-ordered sequential rules common to multiple sequences”. IEEE Transactions on Knowledge and Data Engineering, 27(8), 2203–2216, 2015.
  5. [5] Çelik C, Aladağ Z, Alkan A. “Veri madenciliği ve stok tutma süresi teknikleri ile depolar için dinamik ürün analizi modelinin geliştirilmesi”. Pamukkale Üniversitesi Mühendislik Bilimleri Dergisi, 29(2), 172-182, 2023.
  6. [6] Yıldırım P, Birant D. “Bulut bilişimde veri madenciliği tekniklerinin uygulanması: Bir literatür taraması”. Pamukkale Üniversitesi Mühendislik Bilimleri Dergisi, 24(2), 336-343, 2018.
  7. [7] Hochreiter S, Schmidhuber J. “Long short-term memory”. Neural Computation, 9(8), 1735–1780, 1997.
  8. [8] Aladağ E. “Forecasting of particulate matter with a hybrid ARIMA model based on wavelet transformation and seasonal adjustment”. Urban Climate, 39(1), 100930, 2021.
  9. [9] Md AQ, Kapoor S, Chris Junni AV, Sivaraman AK, Tee KF, Sabireen H, Janakiraman N. “Novel optimization approach for stock price forecasting using multi-layered sequential LSTM”. Applied Soft Computing, 134(1), 1-11, 2023.
  10. [10] Park HJ, Kim Y, Kim HY. “Stock market forecasting using a multi-task approach integrating long short-term memory and the random forest framework”. Applied Soft Computing, 114(1), 1-17, 2022.
  11. [11] Tang P, Tang C, Wang K. “Stock movement prediction: A multi-input LSTM approach”. Journal of Forecasting, 43(5), 1199–1211, 2024.
  12. [12] Shabani M, Tran DT, Magris M, Kanniainen J, Iosifidis A. “Multi-head temporal attention-augmented bilinear network for financial time series prediction”. 30th European Signal Processing Conference (EUSIPCO), Belgrade, Serbia, 29 August-2 September 2022.
  13. [13] Praveen M, Dekka S, Sai DM, Chennamsetty DP, Chinta DP. “Financial time series forecasting: A comprehensive review of signal processing and optimization-driven intelligent models”. Computational Economics, 67(2), 963-989, 2026.
  14. [14] Wang X, Xing Q, Xiao H, Ye M. “Contrastive learning enhanced by graph neural networks for universal multivariate time series representation”. Information Systems, 125(1), 1-13, 2024.
  15. [15] Barazandeh I, Haratizadeh S, Sermpinis G. “A temporal graph-based contrastive approach for financial time series forecasting”. Engineering Applications of Artificial Intelligence, 153(1), 110834, 2025.
  16. [16] Papageorgiou G, Tjortjis C. “Adaptive sliding window normalization”. Information Systems, 129(1), 1-13, 2025.
  17. [17] Liu Q, Tao Z, Tse Y, Wang C. “Stock market prediction with deep learning: The case of China”. Finance Research Letters, 46(1), 1-7, 2022.
  18. [18] Hou Y, Hu C. “Understanding the role of aggregate analyst attention in resolving stock market uncertainty”. Finance Research Letters, 57(1), 1-10, 2023.
  19. [19] Sheng Y, Qu Y, Ma D. “Stock price crash prediction based on multimodal data machine learning models”. Finance Research Letters, 62(1), 1-8, 2024.
  20. [20] Yao D, Yan K. “Time series forecasting of stock market indices based on DLWR-LSTM model”. Finance Research Letters, 68(1), 1-11, 2024.
  21. [21] Kristjanpoller W, Michell K, Llanos C, Minutolo MC. “Incorporating causal notions to forecasting time series: a case study”. Financial Innovation, 11(1), 1-22, 2025.
  22. [22] Niu H, Liu T. “Forecasting the volatility of European Union allowance futures with macroeconomic variables using the GJR-GARCH-MIDAS model”. Empirical Economics, 67(1), 75–96, 2024.
  23. [23] Eroğlu BA, İkizlerli D, Ülkü N. “A mixed-frequency VAR application to studying joint dynamics of foreign investor trading and stock market returns”. Empirical Economics, 67(1), 47–73, 2024.
  24. [24] Li S, Xu S. “Enhancing stock price prediction using GANs and transformer-based attention mechanisms”. Empirical Economics, 68(1), 373–403, 2025.
  25. [25] Özen F, Arslan AE, Beyhan S. “Adaptive sliding window size based sequential learning of LSTM, GRU and transformer models applied to financial time series prediction”. Savunma Bilimleri Dergisi, 22(1), 97-108, 2026.