Petrol Hizmetleri Borsa Yatırım Fonu (ETF) Tahmini için Gramian Açısal Alanın Evrişimsel Sinir Ağı Analizi

Yıl 2025, Cilt: 9 Sayı: 2, 108 - 120, 31.12.2025

Mehmet Sarıkoç , Mete Çelik

https://doi.org/10.33461/uybisbbd.1710520

Öz

Petrol fiyatlarının tahmini, alınacak ekonomik kararlar ve oluşturulacak mali politikalar açısından hem ülkeler hem de şirketler için önemlidir. Ancak finansal fiyat dalgalanmaları doğası gereği doğrusal olmayan, karmaşık ve belirsizdir. Bu nedenlerden dolayı petrol fiyatlarının tahmini zor bir problemdir. Literatürde, petrol fiyatlarını tahmin etmek için istatistiksel ve makine öğrenimi yöntemleri kullanılmıştır. Ancak bu çalışmaların çoğunda petrol fiyatları genellikle zaman serisi olarak temsil edilmiştir. Bu çalışmada, petrol borsa yatırım fonu (ETF) verileri, görüntülerin temsil gücünden faydalanmak için Gramian Açısal Alan (GAF) yöntemi kullanılarak 2 boyutlu görüntü olarak temsil edilmiş ve daha sonra bu görüntü veri kümelerini analiz etmek için AlexNet ve VGG16 evrişimsel sinir ağı (CNN) mimarileri kullanılmıştır. Mevcut ve önerilen GAF-AlexNet ve GAF-VGG16 modellerinin performanslarını test etmek için enerji şirketlerine yatırım yapan bir fon olan VanEck Petrol Hizmetleri ETF'sine (OIH) ait 2016 ve 2022 dönemlerini kapsayan bir veri kümesi kullanılmıştır. Deneysel değerlendirmeler, önerilen modellerin umut verici sonuçlar verdiğini göstermektedir. Bulgular, tahmin modelinin bir ticaret sistemine entegre edilmesinin, araştırmacılara ve yatırımcılara bir karar destek sistemi olarak değerli bilgiler sağlayabileceğini göstermektedir.

Anahtar Kelimeler

: Derin Öğrenme , Zaman Serisi Analizi , ETF , Gramian Açısal Alan (GAA) , Fiyat Tahmini.

Convolutional Neural Network Analysis of the Gramian Angular Field for Oil Services Exchange Traded Fund (ETF) Prediction

Öz

Prediction of oil prices is important for both countries and companies in terms of economic decisions to be made and financial policies to be created. However, due to the nature of financial price fluctuations, they are non-linear, complex, and uncertain. Because of this reasons, prediction of oil prices is a difficult problem. In the literature, statistical and machine learning methods have been used to predict oil prices. However, in most of these studies, oil prices were usually represented as time series. In this study, oil services Exchange-traded fund (ETF) data is represented as a 2D image using Gramian Angular Field (GAF) method, in order to benefit from the representation power of images and then AlexNet and VGG16 convolutional neural network (CNN) architectures are used to analyze this image datasets. To test the performances of existing and the proposed GAF-AlexNet and GAF-VGG16 models, a dataset covering period of 2016 and 2022 belonging to the VanEck Oil Services ETF (OIH), a fund that invests in energy companies, was used. Experimental evaluations show that the proposed models gave promising results. The findings suggest that integrating the predictive model into a trading system can provide valuable insights to researchers and investors as a decision support system.

Anahtar Kelimeler

Deep Learning , Time Series Analysis , ETF , Gramian Angular Field (GAF) , Price Prediction.

Etik Beyan

It is declared that during the preparation process of this study, scientific and ethical principles were followed, and all the studies benefited from are stated in the bibliography.

Kaynakça

• Abd Elaziz, M., Ewees, A. A., & Alameer, Z. (2020). Improving adaptive neuro-fuzzy inference system based on a modified salp swarm algorithm using genetic algorithm to forecast crude oil price. Natural Resources Research, 29, 2671–2686.
• Abdollahi, H., & Ebrahimi, S. B. (2020). A new hybrid model for forecasting Brent crude oil price. Energy, 200, 117520.
• Arratia, A., & Eduardo, S. (2020). Convolutional neural networks, image recognition and financial time series forecasting. In Mining Data for Financial Applications (pp. 60–69). Springer. https://doi.org/10.1007/978-3-030-37720-5_5
• Barra, S., Carta, S. M., Corriga, A., Podda, A. S., & Recupero, D. R. (2020). Deep learning and time series-to-image encoding for financial forecasting. IEEE/CAA Journal of Automatica Sinica, 7, 683–692.
• Barunik, J., & Malinska, B. (2016). Forecasting the term structure of crude oil futures prices with neural networks. Applied Energy, 164, 366–379.
• Chauhan, J. K., Ahmed, T., & Sinha, A. (2023, December). Comparative Analysis of CNN Pre-trained Models for Stock Market Trend Prediction. In International Conference on Recent Trends in Image Processing and Pattern Recognition (pp. 110–129). Springer Nature Switzerland.
• Chen, J. H., & Tsai, Y. C. (2020). Encoding candlesticks as images for pattern classification using convolutional neural networks. Financial Innovation, 6(1), 1–19.
• Demirezen, M. U., Civrizoğlu, A., & Yavanoğlu, U. (2021). Sualtı objelerinin makine öğrenmesi yöntemleri ile tespitinde zaman serisi-görüntü dönüşümü tabanlı yeni yaklaşımlar. Gazi Üniversitesi Mühendislik Mimarlık Fakültesi Dergisi, 36(3), 1399-1416. https://doi.org/10.17341/gazimmfd.826453
• Estebsari, A., & Rajabi, R. (2020). Single residential load forecasting using deep learning and image encoding techniques. Electronics, 9(1), 68.
• Estebsari, A., & Rajabi, R. (2020). Single Residential Load Forecasting Using Deep Learning and Image Encoding Techniques. Electronics, 9(1), 68. https://doi.org/10.3390/electronics9010068
• Fan, L., Pan, S., Li, Z., & Li, H. (2016). An ICA-based support vector regression scheme for forecasting crude oil prices. Technological Forecasting and Social Change, 112, 245–253.
• Hou, A., & Suardi, S. (2012). A nonparametric GARCH model of crude oil price return volatility. Energy Economics, 34(2), 618–626.
• Kibot. (2023, May). Free historical data. http://www.kibot.com/free_historical_data.aspx
• Kilicarslan, S., Celik, M., & Sahin, Ş. (2021). Hybrid models based on genetic algorithm and deep learning algorithms for nutritional anemia disease classification. Biomedical Signal Processing and Control, 63, 102231.
• Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2012). ImageNet classification with deep convolutional neural networks. In Proceedings of the 25th International Conference on Neural Information Processing Systems (pp. 1097–1105). Lake Tahoe, NV.
• LeCun, Y. (2023, April 30). Lenet-5, convolutional neural networks. http://yann.lecun.com/exdb/lenet/
• Lertthaweedech, W., Kantavat, P., & Kijsirikul, B. (2022). Effective crude oil trading techniques using long short-term memory and convolution neural networks. Journal of Advances in Information Technology, 13(6).
• Liu, S., Fang, W., Gao, X., An, F., Jiang, M., & Li, Y. (2019). Long-term memory dynamics of crude oil price spread in nondollar countries under the influence of exchange rates. Energy, 182, 753–764.
• Long, H. W., Ho, O. I., He, Q. Q., & Si, Y. W. (2025). Transfer Learning in Financial Time Series with Gramian Angular Field. arXiv preprint arXiv:2504.00378.
• Moshiri, S., & Foroutan, F. (2006). Forecasting nonlinear crude oil futures prices. The Energy Journal, 27(4).
• Naftali, C., Tucker, B., & Manuela, V. (2021). Trading via image classification. In Proceedings of the First ACM International Conference on AI in Finance (ICAIF '20), Article 53, 1–6.
• Ozkok, F. O., & Celik, M. (2021). Convolutional neural network analysis of recurrence plots for high resolution melting classification. Computational Methods and Programs in Biomedicine, 207, 106139.
• Ozkok, F. O., & Celik, M. (2023). Classification of high resolution melting curves using recurrence quantification analysis and data mining algorithms. In Springer Lecture Notes on Data Engineering and Communications Technologies (pp. 641–650). Springer.
• Paheding, S., Reyes, A. A., Kasaragod, A., & Oommen, T. (2022). GAF-NAU: Gramian angular field encoded neighborhood attention U-Net for pixel-wise hyperspectral image classification. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition (pp. 409–417).
• Pandey, V. K., Ahmed, T., & Sahoo, G. (2025, January). Comparative Analysis of Stock Price Forecasting using MTF Image Transformation: Evaluation of Random Forest and CNN Models. In 2025 International Conference on Intelligent Systems and Computational Networks (ICISCN) (pp. 1–6). IEEE
• Pumi, G., Valk, M., Bisognin, C., Bayer, F. M., & Prass, T. S. (2019). Beta autoregressive fractionally integrated moving average models. Journal of Statistical Planning and Inference, 200, 196–212.
• Ramyar, S., & Kianfar, F. (2017). Forecasting crude oil prices: A comparison between artificial neural networks and vector autoregressive models. Computational Economics, 53(2), 743–761.
• Rather, A. M., Agarwal, A., & Sastry, V. N. (2015). Recurrent neural network and a hybrid model for prediction of stock returns. Expert Systems with Applications, 42(6), 3234–3241.
• Shahid, S. M., Ko, S., & Kwon, S. (2022). Performance Comparison of 1D and 2D Convolutional Neural Networks for Real-Time Classification of Time Series Sensor Data. 2022 International Conference on Information Networking (ICOIN) (pp.507–511). https://doi.org/10.1109/ICOIN53446.2022.9687284.
• Salamai, A. A. (2023). Deep learning framework for predictive modeling of crude oil price for sustainable management in oil markets. Expert Systems with Applications, 211, 118658.
• Sarıkoç, M., & Celik, M. (2025). PCA-ICA-LSTM: A Hybrid Deep Learning Model Based on Dimension Reduction Methods to Predict S&P 500 Index Price. Comput Econ 65, 2249–2315. https://doi.org/10.1007/s10614-024-10629-x.
• Sezer, O. B., & Ozbayoglu, A. M. (2018). Algorithmic financial trading with deep convolutional neural networks: Time series to image conversion approach. Applied Soft Computing, 70, 525–538.
• Simonyan, K., & Zisserman, A. (2014). Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:1409.1556.
• Srinivasamurthy, R. S. (2018). Understanding 1D convolutional neural networks using multiclass time-varying signals (Master's thesis, Clemson University).
• Thakkar, A., & Chaudhari, C. (2021). A comprehensive survey on deep neural networks for stock market: The need, challenges, and future directions. Expert Systems with Applications, 177, 114800.
• Wang, Z., & Oates, T. (2015). Encoding time series as images for visual inspection and classification using tiled convolutional neural networks. In Workshops at the Twenty-Ninth AAAI Conference on Artificial Intelligence (Vol. 1). Menlo Park, CA, USA: AAAI.
• Wu, J., Zhang, Z., Tong, R., Zhou, Y., Hu, Z., & Liu, K. (2023). Imaging feature-based clustering of financial time series. PLOS ONE, 18(7), e0288836.
• Wu, Y., Yang, F., Liu, Y., Zha, X., & Yuan, S. (2018). A comparison of 1-D and 2-D deep convolutional neural networks in ECG classification. arXiv preprint arXiv:1810.07088.
• Xie, W., Yu, L., & Xu, S. (2006). New method for crude oil price forecasting based on support vector machines. In Proceedings of the 6th International Conference on Computational Science (pp. 444–451).
• Yu, L., Wang, S., & Lai, K. K. (2008). Forecasting crude oil price with an EMD-based neural network ensemble learning paradigm. Energy Economics, 30(5), 2623–2635.
• Zhang, S., Luo, J., Wang, S., & Liu, F. (2023). Oil price forecasting: A hybrid GRU neural network based on decomposition–reconstruction methods. Expert Systems with Applications, 218, 119617.
• Zhao, C., & Wang, B. (2014). Forecasting crude oil price with an autoregressive integrated moving average (ARIMA) model. In Fuzzy Information & Engineering and Operations Research & Management (pp. 275–286). Springer.