시계열 분해와 하이브리드 순환신경망 기반 스프레드 예측: 국제 에너지 선물 시장을 중심으로
Time Series Decomposition and Hybrid RNN-Based Spread Prediction: Focused on the International Energy Futures Market
0
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
에너지 선물 스프레드는 안정적 위험 헤지, 수익성 있는 선물 투자, 에너지 및 주식 시장에 대한 신호로써의 유용성 때문에 정확한 스프레드 예측은 여러 시장 참여자들에게 위험관리 및 수익 극대화 차원에서 중요한 과제이다. 따라서 본 연구는 에너지 선물 스프레드에 내포된 여러 가지 시계열적 특성을 고려하여 미래 스프레드 예측력 향상을 목표로 하였다. 이에 본 연구는 RobustSTL을 사용하여 스프레드를 시계열 요소 즉, 추세, 계절성, 노이즈로 나누고, 각 요소의 특성에 따라 ARIMA와 RNN(LSTM, GRU)을 사용하여 요소별로 예측한 후 합산하여 미래 스프레드를 예측하는 STAR-RNN 모형을 제시한다. 본 연구에서는 2005년부터 2021년까지의 WTI 원유, 브렌트유, 난방유, 천연가스 선물로 이루어진 5개의 스프레드 쌍 데이터를 사용하여 실험하였다. 실험 결과, STAR-RNN 중 STAR-LSTM은 평균적으로 LSTM 대비 RMSPE 기준 6.8%, ARIMA 대비 8.6%의 예측 오차가 감소했고, STAR-GRU는 GRU 대비 10.4%, ARIMA 대비 9.9%의 오차가 감소했다. 이러한 결과는 원 시계열로부터 추세, 계절성, 노이즈를 분리하여 예측하는 방식이 최종 스프레드 예측 성능 향상에 있어 중요한 역할을 한다는 것을 의미한다.
Abstract
Since futures spreads are used strategically as a signal for market direction or as an instrument to hedge against risk or make profits, accurate spread prediction is crucial to many market participants who are willing to manage risk or capitalize on profitable opportunities. Therefore, this study aims to improve the predictive power of future spreads by considering various time-series characteristics embedded in energy futures spreads. In this regard, we present STAR-RNN, a spread prediction model that combines RobustSTL, ARIMA, and RNN (LSTM and GRU), by decomposing spreads into time series elements, trends, seasonalities, and noise, to predict future time series. The experiment was conducted using spread data calculated from WTI crude oil, Brent oil, heating oil, and natural gas futures from 2005 to 2021. As a result of the experiment, STAR-LSTM reduced RMSPE by 6.8% and 8.6% on average, compared to LSTM and ARIMA, and STAR-GRU reduced RMSPE by 10.4% and 9.9% on average, compared to GRU and ARIMA. These results indicate that the decomposed elements from the original time series play an important role in improving the final spread prediction performance.
참고문헌 (45)
- 고봉찬, "주가 추세요인의 추출방법과 수익성 분석", 재무관리연구, 38(2), pp.251~288 (2021), 한국재무관리학회[DOI]
- 김범수, "페어트레이딩 전략의 수익성 연구 : 해외 선물시장을 중심으로", 경영과학, 33(4), pp.1~15 (2016), 한국경영과학회[DOI]
- 김상배, "비대칭 오차수정모형을 활용한 유가와 주가지수의 관계 분석: 유가증권시장을 중심으로", 재무관리연구, 36(3), pp.141~168 (2019), 한국재무관리학회[DOI]
- 김연규, "제1장: 셰일혁명과 신국제 에너지 질서", 세계정치, 23, pp.29~81 (2015)
- 김영민, "주가지수 선물의 가격 비율에 기반한 차익거래 투자전략을 위한 페어트레이딩 규칙 개발", 한국산업경영시스템학회지, 37(4), pp.202~211 (2014), 한국산업경영시스템학회
- ETS를 이용한 시계열자료분석 I, "SAS", pp.2002, 박유성
- 옥철, "WHO, 코로나19 팬데믹 선언", 연합뉴스, ~ (2020)
- 이상열, "시계열 분석 이론 및 SAS 실습" (2013), 자유아카데미
- 정인곤, "구조적 변화 감지 과정이 포함된 페어트레이딩 알고리즘의 성과분석", 한국경영과학회지, 42(3), pp.13~24 (2017), 한국경영과학회[DOI]
- 진우성, "리포트-제 2 의 셰일혁명은 무엇이 다른가?", Korea Petroleum Association Journal, pp.38~41 (2017)
- Akiba, T., "Optuna: A Next-Generation Hyperparameter Optimization Framework", Proceedings of the 25th ACM KDD International Conference on Knowledge Discovery & Data Mining (2019)
- Bao, W., "A Deep Learning Framework for Financial Time Series Using Stacked Autoencoders and Long-Short Term Memory", Plos One, 12(7), pp.1~24 (2017)
- Bibby, J., "Prediction and Improved Estimation in Linear Models" (1977), Wiley
- Brim, A., "Deep Reinforcement Learning Pairs Trading with a Double Deep Q-Network", 2020 10th Annual Computing and Communication Workshop and Conference (2020), IEEE
- Borowski, K., "Analysis of Selected Seasonality Effects in Markets of Futures Contracts With the Following Underlying Instruments: Crude Oil, Brent Oil, Heating Oil, Gas Oil, Natural Gas, Feeder Cattle, Live Cattle, Lean Hogs and Lumber", Journal of Management and Financial Sciences, 9(26), pp.27~44 (2016)
- Buyuksahin, B., "Physical Markets, Paper Markets and the WTI-Brent Spread", The Energy Journal, 34(3), pp.129~151 (2013)
- Cahuantzi, R., "A Comparison of LSTM and GRU Networks for Learning Symbolic Sequences", ~ (2021)
- Choi, H. K., "Stock Price Correlation Coefficient Prediction with ARIMA-LSTM Hybrid Model", ~ (2018)
- Cho, K., "Learning Phrase Representations Using RNN Encoder-Decoder for Statistical Machine Translation", ~ (2014)
- Cleveland, R. B., "STL: A Seasonal-Trend Decomposition Procedure Based on Loess (With Discussion)", Journal of Official Statistics, 6, pp.3~73 (1990)
- "Introduction to Crack Spreads" (2017)[URL]
- De Myttenaere, A., "Mean absolute percentage error for regression models", Neurocomputing, 192, pp.38~48 (2016)
- Dickey, D. A, "Distribution of the Estimators for Autoregressive Time Series With a Unit Root", Journal of the American Statistical Association, pp.427~431 (1979)
- Dunis, C. L., "Trading Futures Spreads: An Application of Correlation and Threshold Filters", Applied Financial Economics, 16, pp.903~914 (2006)
- "International Crude Oil Market Handbook" (2011)[URL]
- Gatev, P., "Feedforward Ankle Strategy of Balance during Quiet Stance in Adults", The Journal of Physiology, 514(3), pp.915~928 (1999)
- Gatev, E., "Pairs Trading: Performance of a Relative-Value Arbitrage Rule", The Review of Financial Studies, 19(3), pp.797~827 (2006)
- Géron, A., "Hands-on Machine Learning with Scikit-Learn, Keras, and Tensorflow:Concepts, Tools, and Techniques to Build Intelligent Systems" (2019), O’Reilly Media, Inc
- Hamzaçebi, C., "Improving Artificial Neural Networks’ Performance in Seasonal Time Series Forecasting", Information Sciences, 178(23), pp.4550~4559 (2008)
- Hochreiter, S., "Long Short-Term Memory", Neural Computation, 9(8), pp.1735~1780 (1997)
- Holt, C. C, "Forecasting Seasonals and Trends by Exponentially Weighted Moving Averages", Journal of American Statistical Association, pp.112~125 (1957)
- "Hudson and Thames"[URL]
- Hyndman, R. J, "Forecasting: Principles and Practice" (2018), Otexts
- Ji, L, "Carbon Futures Price Forecasting Based With ARIMA-CNN-LSTM Model", Procedia Computer Science, pp.33~38 (2019)
- Khandelwal, I., "Time Series Forecasting Using Hybrid ARIMA and ANN Models Based on DWT Decomposition", Procedia Computer Science, 48, pp.173~179 (2015)
- Kim, S. -H., "Hybrid Deep Reinforcement Learning for Pairs Trading", Applied Sciences, 12(3), pp.944~ (2022)
- Mangalampalli, R., "Stock Prediction Using Hybrid ARIMA and GRU Models", International Journal of Engineering Research & Technology, 9(5), pp.2278~0181 (2020)
- Milonas, N. T., "Price Spread and Convenience Yield Behaviour in the International Oil Market", Applied Financial Economics, 11(1), pp.23~36 (2001)
- Patro, S., "Normalization: A Preprocessing Stage", ~ (2015)
- Smith, T. G, "pmdarima: ARIMA Estimators for Python", ~ (2017)
- Song, D., "Forecasting Stock Market Indices Using Padding-Based Fourier Transform Denoising and Time Series Deep Learning Models", IEEE Access, 9, pp.83786~83796 (2021)
- Suna, S., "Seasonal and Trend Forecasting of Tourist Arrivals: An Adaptive Multiscale Ensemble Learning Approach", ~ (2020)
- Tang, Q., "Prediction of Financial Time Series Using LSTM and Data Denoising Methods", ~ (2021)
- Wang, H., "Robust Regression Shrinkage and Consistent Variable Selection Through the Lad-Lasso", Journal of Business & Economic Statistics, 25(3), pp.347~355 (2007)
- Wen, Q, "RobustSTL: A Robust Seasonal-Trend Decomposition Algorithm for Long Time Series", Proceedings of the AAAI Conference on Artificial Intelligence, 33(01) (2019)