Comparison of SARIMA and Hybrid SARIMA-SVR Models for Forecasting Water Distribution Volume in Mueang District, Chachoengsao Province
Article Sidebar
Main Article Content
Abstract
This research aimed to study and compare the efficiency of forecasting models for water distribution volume in Chachoengsao Province, an economic zone experiencing continuous growth in water demand. The study compared between a standalone Seasonal Autoregressive Integrated Moving Average (SARIMA) model and a hybrid model combining SARIMA with Support Vector Regression (SARIMA-SVR). The dataset used in this study comprised monthly water distribution volumes spanning 132 months, from January 2015 to December 2025. This dataset was partitioned into two subsets: a training dataset of 120 months for model development and a testing dataset of 12 months for out-of-sample forecast validation. Predictive performance was evaluated using Mean Absolute Error (MAE) and Mean Absolute Percentage Error (MAPE). The results revealed that the SARIMA-SVR2 hybrid model provided the highest accuracy in forecasting water distribution volume under the studied data and model, with a MAE of 20,260.07 cubic meters and a MAPE of 2.35%, achieving a forecasting accuracy of 97.65%. This superior performance is attributed to the SARIMA component's ability to effectively capture linear trends of seasonal patterns and to SVR component’s ability to efficiently minimize errors derived from nonlinear residuals. The findings of this study can serve as a guideline to support decision-making for relevant agencies in planning and managing water reserves and developing public utility infrastructure in Mueang District, Chachoengsao Province.
Article Details
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
บทความที่ได้รับการตีพิมพ์เป็นลิขสิทธิ์ของ วารสารวิทยาศาสตร์และเทคโนโลยี มหาวิทยาลัยอุบลราชธานี
ข้อความที่ปรากฏในบทความแต่ละเรื่องในวารสารวิชาการเล่มนี้เป็นความคิดเห็นส่วนตัวของผู้เขียนแต่ละท่านไม่เกี่ยวข้องกับมหาวิทยาลัยอุบลราชธานี และคณาจารย์ท่านอื่นๆในมหาวิทยาลัยฯ แต่อย่างใด ความรับผิดชอบองค์ประกอบทั้งหมดของบทความแต่ละเรื่องเป็นของผู้เขียนแต่ละท่าน หากมีความผิดพลาดใดๆ ผู้เขียนแต่ละท่านจะรับผิดชอบบทความของตนเองแต่ผู้เดียว
References
Zhang, G.P. 2003. Time series forecasting using a hybrid ARIMA and neural network model. Neurocomputing. 50: 159-175.
Box, G.E.P. and et al. 2008. Time Series Analysis: Forecasting and Control. 4th edition. Hoboken: John Wiley and Sons, Inc.
Pai, P.F. and Lin, C.S. 2005. A hybrid ARIMA and support vector machines model in stock price forecasting. Omega. 33(6): 497-505.
Sutthison, T. 2020. Appropriate models for forecasting of water supply consumption in Ubonratchathani Province. Burapha Science Journal. 25(1): 28-50. (in Thai)
Albarr, H. and Kusumawati, R. 2023. Hybrid autoregressive integrated moving average-support vector regression for stock price forecasting. Jurnal Matematika, Sains, dan Teknologi. 24(2): 1-17.
Provincial Waterworks Authority. 2025. Data Water Consumption. https://www.pwa.co.th/province/. Accessed 20 February 2026. (in Thai)
Bowerman, B.L., O’Connell, R.T. and Koehler, A.B. 2005. Forecasting, Time series and Regression: An Applied Approach. 4th edition. Belmont: Thomson Brooks/Cole.
Vapnik, V.N. 1995. The Nature of Statistical Learning Theory. Berlin: Springer-Verlag.
Ojemakinde, B.T. 2006. Support Vector Regression for Non-Stationary Time Series. M.Sc. Thesis, University of Tennessee, Knoxville.
Scholkopf, B. and Smola, A.J. 2001. Learning with Kernels: Support Vector Machines, Regularization, Optimization and Beyond. Cambridge: MIT Press.
Awad, M. and Khanna, R. 2015. Efficient Learning Machines. Berkeley: Apress.
Meyer, D. and et al. 2025. e1071: Misc functions of the Department of Statistics, Probability Theory Group (Formerly: E1071), TU Wien. https://CRAN.R-project.org/package=e1071. Accessed 22 April 2026.
Lewis, C.D. 1982. Industrial and Business Forecasting Methods: A Practical Guide to Exponential Smoothing and Curve Fitting. Oxford: Butterworth-Heinemann.