Abstract
Amaç- Bu çalışmanın amacı, Makine Öğrenmesi yöntemlerinden yararlanarak geliştirilen modellerin zaman serilerinin öngörüsünde alternatif bir yöntem olup olmadığının incelenmesidir.
Yöntem- Geleneksel olarak, Otoregresif Entegre Hareketli Ortalama (ARIMA) modeli, zaman serisi tahmininde en yaygın kullanılan doğrusal modellerden biridir. Çalışmada,ARIMA modellerinin yanı sıra Rassal Orman ve Hibrit Rassal Orman yöntemleri kullanılmış ve Türkiye Konut Fiyat Endeksi serisi için bu modellerin öngörü performansları karşılaştırılmıştır.
Bulgular- Hibrit modelin konut fiyat endeksini öngörmede diğer yöntemlerden daha başarılı olduğu tespit edilmiştir.
Sonuç- Sonuç olarak, ARIMA ve Makine Öğrenmesi yöntemini birleştiren hibrit modellerin, ekonomik ve finansal verilerin öngörüsünde alternatif bir yöntem olarak kullanılabileceği tespit edilmiştir.
References
- Franses, P. H., & Van Dijk, D. (2003). Non-linear Time Series Models in Empirical Finance. Cambridge University Press.
- Hastie, T., Tibshirani, R., &Friedman, J. (2009). The Elements of Statistical Learning. New York: Springer-Verlag.
- Kumar, M., &Thenmozhi, M. (2014). Forecasting Stock Index Returns using ARIMA-SVM, ARIMA-ANN, and ARIMA-Random Forest Hybrid Models. Int. J. Banking, Accounting and Finance, 284 – 308.
- Pai, P-F., & Lin, C-S. (2005). A Hybrid ARIMA and Support Vector Machines Model in Stock Price Forecasting. Omega, 497 – 505.
- Pedregosa, F.(2016). Hyperparameter Optimization With Approximate Gradient. JMLR: W&CP vol.48, 737-746.
- TCMB(2019, 27 Ekim). KonutFiyatEndeksi. https://www.tcmb.gov.tr/wps/wcm/connect/b4628fa9-11a7-4426-aee6-dae67fc56200/KFE-Metaveri.pdf?MOD=AJPERES&CACHEID=ROOTWORKSPACE-b4628fa9-11a7-4426-aee6-dae67fc56200-mDXEz4N
- Wang, P. (2008). Financial Econometrics. Routledge.
- Zhang, G. P. (2003). Time Series Forecasting using A Hybrid ARIMA and Neural Network Model. Neurocomputing, 159 – 175.
- Zhang, G., Patuwo, B. E., & Hu M. Y. (1998). Forecasting with Artificial Neural Networks: The State of The Art. International Journal of Forecasting, 35–62
Abstract
Purpose- The aim of this study is to investigate whether the models developed by using Machine Learning methods are an alternative method for forecasting time series.
Methodology-Traditionally, the Autoregressive Integrated Moving Average (ARIMA) model has been one of the most widely used linear models in time series forecasting. In the study, we use Random Forest and Hybrid Random Forest-ARIMA models besides the ARIMA model and compare their forecasting performance for the Turkish Housing Price Index series.
Findings- The hybrid model was found to be more successful than other methods in forecasting the housing price index.
Conclusion- As a result, hybrid models that combine ARIMA and machine learning method can be used an alternative method in forecasting economic and financial data
References
- Franses, P. H., & Van Dijk, D. (2003). Non-linear Time Series Models in Empirical Finance. Cambridge University Press.
- Hastie, T., Tibshirani, R., &Friedman, J. (2009). The Elements of Statistical Learning. New York: Springer-Verlag.
- Kumar, M., &Thenmozhi, M. (2014). Forecasting Stock Index Returns using ARIMA-SVM, ARIMA-ANN, and ARIMA-Random Forest Hybrid Models. Int. J. Banking, Accounting and Finance, 284 – 308.
- Pai, P-F., & Lin, C-S. (2005). A Hybrid ARIMA and Support Vector Machines Model in Stock Price Forecasting. Omega, 497 – 505.
- Pedregosa, F.(2016). Hyperparameter Optimization With Approximate Gradient. JMLR: W&CP vol.48, 737-746.
- TCMB(2019, 27 Ekim). KonutFiyatEndeksi. https://www.tcmb.gov.tr/wps/wcm/connect/b4628fa9-11a7-4426-aee6-dae67fc56200/KFE-Metaveri.pdf?MOD=AJPERES&CACHEID=ROOTWORKSPACE-b4628fa9-11a7-4426-aee6-dae67fc56200-mDXEz4N
- Wang, P. (2008). Financial Econometrics. Routledge.
- Zhang, G. P. (2003). Time Series Forecasting using A Hybrid ARIMA and Neural Network Model. Neurocomputing, 159 – 175.
- Zhang, G., Patuwo, B. E., & Hu M. Y. (1998). Forecasting with Artificial Neural Networks: The State of The Art. International Journal of Forecasting, 35–62
Details
| Primary Language | Turkish |
|---|---|
| Subjects | Finance, Business Administration |
| Journal Section | Articles |
| Authors | |
| Publication Date | December 30, 2019 |
| Published in Issue | Year 2019 Volume: 10 Issue: 1 |
Cite
PressAcademia Procedia (PAP) publishes proceedings of conferences, seminars and symposiums. PressAcademia Procedia aims to provide a source for academic researchers, practitioners and policy makers in the area of social and behavioral sciences, and engineering.
PressAcademia Procedia invites academic conferences for publishing their proceedings with a review of editorial board. Since PressAcademia Procedia is an double blind peer-reviewed open-access book, the manuscripts presented in the conferences can easily be reached by numerous researchers. Hence, PressAcademia Procedia increases the value of your conference for your participants.
PressAcademia Procedia provides an ISBN for each Conference Proceeding Book and a DOI number for each manuscript published in this book.
PressAcademia Procedia is currently indexed by DRJI, J-Gate, International Scientific Indexing, ISRA, Root Indexing, SOBIAD, Scope, EuroPub, Journal Factor Indexing and InfoBase Indexing.
Please contact to procedia@pressacademia.org for your conference proceedings.