Applied integration of time series and multi-variable regression algorithms

Abstract

Time Series (TS) based prediction models generate prediction based data that is supposed to be similar to the future data at a certain level. In this study, we designed new modeling that increases the prediction performance of the TS algorithm. The main purpose of the new modeling is to integrate the Multivariate-Adaptive-Regression-Splines (MARSplines) algorithm into the TS algorithm. Five-year Tokyo Stock Exchange data is analyzed as a case study to apply the relevant models. The results show that the new regression-based approach significantly improves the prediction performance of the time series algorithm.

Keywords

• Exponential correction
• MARS
• multivariate adaptive regression
• stock price
• time series analysis.

Applied integration of time series and multi-variable regression algorithms

Abstract

Zaman Serisi (ZS) tabanlı tahmin modelleri, belirli bir düzeyde geçmiş verilere benzer fonksiyonel dağılıma sahip, tahmine dayalı veriler üretir. Bu çalışmada, ZS algoritmasının tahmin performansını artıran yeni bir modelleme tasarlanmıştır. Yeni modellemenin temel amacı, Çok Değişkenli Uyarlamalı Regresyon Katmanları (MARSplines) algoritmasını ZS algoritmasına entegre etmektir. Beş yıllık Tokyo Menkul Kıymetler Borsası verileri, ilgili modelleri uygulamak için bir vaka çalışması olarak analiz edilmiştir. Sonuçlar, yeni regresyon temelli yaklaşımın ZS algoritmasının tahmin performansını önemli ölçüde geliştirdiğini göstermiştir. 

Keywords

• Time series analysis
• multivariate adaptive regression
• stock price
• exponential correction

References

1. A. Guolo, C. Varin. (2014). Beta Regression For Time Series Analysis Of Bounded Data, With Application To Canada Google R Flu Trends, The Annals of Applied Statistics, Institute of Mathematical Statistics, Vol. 8, No. 1, 74–88, Doi: 10.1214/13-AOAS684
2. S. Arasu, M. Jeevananthan, N. Thamaraiselvan, B. Janarthanan. (2014). Performances of data mining techniques in forecasting stock index – evidence from India and US, J.Natn.Sci.Foundation Sri Lanka 42 (2): 177–191, DOI: http://dx.doi.org/10.4038/jnsfsr.v42i2.6989
3. J. H. Friedman. (1990). Multivariate Adaptive Regression Splines , Dept. of Statistics Tech. Report 102
4. J. R. Leathwick, D. Rowe, J. Richardson, J. Elith, T. Hastie. (2005). Using multivariate adaptive regression splines to predict the distributions of New Zealand’s freshwater diadromous fish, Freshwater Biology 50, 2034–2052, doi:10.1111/j.1365-2427.2005.01448.x
5. M. M. Al-Idrisi. (1991). Use of Regression and Triple Exponential Smoothing Models for Forecasting Share Prices of Saudi Companies, JKAU: Econ. & Adm. vol. 4, pp. 3-25 (1411 A.H. / 1991 A.D.)
6. M. C. A. Neto, G. Tavares, V. M. O. Alves, G. D. C. Cavalcanti, T. I. Ren. (2010). Improving financial time series prediction using exogenous series and neural networks committees , The 2010 International Joint Conference on Neural Networks (IJCNN), Barcelona, 2010, pp. 1-8., doi: 10.1109/IJCNN.2010.5596911
7. P. S. Kalekar, Time series Forecasting using Holt-Winters Exponential Smoothing (2004), pp. 2-3
8. Electronic Statistics Textbook (1995), http://www.statsoft.com/Textbook/Multivariate-Adaptive-Regression-Splines

9. T. B. Fomby, Exponential Smoothing Models (2008), http://faculty.smu.edu/tfomby/eco5375/data/SMOOTHING%20MODELS_V6.pdf, pp. 9-10
10. O. Kisi, K. S. Parmar. (2016) Application of least square support vector machine and multivariate adaptive regression spline models in long term prediction of river water pollution. J Hydrol 534:104–112]
11. O. Hamidi, L. Tapak, H. Abbasi, Z. Maryanaji. (2017). Application of random forest time series, support vector regression and multivariate adaptive regression splines models in prediction of snowfall (a case study of Alvand in the middle Zagros, Iran). Theoretical and Applied Climatology. 134. 1-8. 10.1007/s00704-017-2300-9.
12. L. J. Kao, C. C. Chiu. (2020). Application of integrated recurrent neural network with multivariate adaptive regression splines on SPC-EPC process, Journal of Manufacturing Systems, Vol. 57, Pages 109-118, ISSN 0278-6125, doi.org/10.1016/j.jmsy.2020.07.020.
13. C. Grillenzoni, M. Fornaciari. (2019). On-line peak detection in medical time series with adaptive regression methods. Econometrics and Statistics, Vol. 10, Pages 134-150, ISSN 2452-3062, doi.org/10.1016/j.ecosta.2018.07.002.
14. S. Lee, S. Lee, M. Moon. (2020). Hybrid change point detection for time series via support vector regression and CUSUM method. Applied Soft Computing, Vol. 89, 106101, ISSN 1568-4946, doi.org/10.1016/j.asoc.2020.106101.
15. G. R. Mode, K. A. Hoque. (2020). Adversarial Examples in Deep Learning for Multivariate Time Series Regression. eprint=2009.11911, arXiv, cs.LG
16. Y. Okkaoglu, Y. Akdi, , E. Golveren, M. Yucel. (2020). Estimation and forecasting of PM10 air pollution in Ankara via time series and harmonic regressions. International Journal of Environmental Science and Technology.(doi:10.1007/s13762-020-02705-0)
17. S. Jiang. (2019)."Combining Deep Neural Networks and Classical Time Series Regression Models for Forecasting Patient Flows in Hong Kong," in IEEE Access, vol. 7, pp. 118965-118974, doi: 10.1109/ACCESS.2019.2936550.
18. I. Ilic, B. Gorgulu, M. Cevik, M. G. Baydogan. (2020).Explainable boosted linear regression for time series forecasting. eprint=2009.09110, arXiv, cs.LG