Improving Prediction Accuracy of Time Series Data Using ARIMA-ANN Hybrid Model and Regression Analysis

Main Article Content

นัท กุลวานิช


The new hybrid model using linear regression methods to combine both ARIMA and Artificial Neural Networks (ARIMA-ANN-REG Hybrid Model) achieves better prediction accuracy than the traditional hybrid model combining ARIMA and ANN (ARIMA-ANN Hybrid Model). This research examines and compares the prediction accuracy among ARIMA model, ARIMA-ANN model, and ARIMA-ANN-REG model by using both real and simulated data for the comparison of prediction accuracy. The simulated data are generated from 8 ARIMA processes and the real data are comprised of six real datasets in Thailand. The results suggest that ARIMA-ANN-REG model has the highest prediction accuracy in both real and simulated data.


Download data is not yet available.

Article Details

Author Biography

นัท กุลวานิช, Chulalongkorn University

ภาควิชาสถิติ คณะพาณิชยศาสตร์และการบัญชี จุฬาลงกรณ์มหาวิทยาลัย แขวงวังใหม่ เขตปทุมวัน กรุงเทพมหานคร 10330


[1] Taesombat, S., 2010, Quantitative Forecasting, Kasetsart University Press, Bangkok, 487 p. (in Thai)
[2] Shumway, R.H. and Stoffer, D.S., 2010, Time Series Analysis and Its Applications with R Examples, 3rd Ed., Springer, Berlin, 576 p.
[3] Zhang, P.G., 2003, Time series forecasting using a hybrid ARIMA and neural network model, Neurocomputing 50: 159-175.
[4] Faruk, D.O., 2010, A hybrid neural network and ARIMA model for water quality time series prediction, Eng. Appl. Artif. Intell. 23: 586-594.
[5] Meth, N., Saxena, V.P. and Pardasani, K.R., 2010, A comparison between hybrid approaches of ANN and ARIMA for Indian stock trend forecasting, Bus. Intell. J. 3: 23-43.
[6] Koutroumanidis, T., Ioannou, K. and Arabatzis, G., 2009, Predicting fuel wood prices in Greece with the use of ARIMA models artificial neural networks and a hybrid ARIMA-ANN model, Energy Pol. 37: 3627-3634.
[7] Ebrahimi, A., 2019, Time series forecasting of styrene price using a hybrid ARIMA and neural network model, Indep. J. Manag. Prod. 10: 915-933.
[8] Siripanich, P., Nillaporn, P. and Trakarnta lerngsuk, S., 2007, Time Series Forecasting Using a Combined ARIMA and Artificial Neural Network Model of Styrene Price Using a Hybrid ARIMA and Neural Network Model, Operations Research Network 2007, 7 p. (in Thai)
[9] Boonmana, C. and Kulvanich, N., 2017, A comparative prediction accuracy of hybrid time series models, Thai Sci. Technol. J. 25(2): 177-190. (in Thai)
[10] Somsila, C., Chiewchanwattana, S. and Sunat, K., 2010, Hybrid Model for Air Quality Data Prediction Case Study: Air Quality Data in Thailand, The 11th Graduate Research Conference, 12 p. (in Thai)
[11] Khairalla, M., Xu-Ning and AL-Jallad, N.T., 2017, Hybrid forecasting scheme for financial time-series data using neural network and statistical methods, IJACSA 8: 319-327.
[12] Zhang, P.G., 2007, A neural network ensemble method with jittered training data for time series forecasting, Inform. Sci. 177: 5329-5346.