Research article

Forecasting the total electricity production in South Africa: Comparative analysis to improve the predictive modelling accuracy

  • Received: 24 December 2018 Accepted: 28 January 2019 Published: 31 January 2019
  • Electricity plays an important role in the South African economy with the industrial sector consuming the highest proportion followed by the residential and mining sector. Besides the fact that electricity is considered as an important energy sources, an adequate supply of electricity remains an important factor that affects the development and economic growth of a country. Therefore, it becomes even more important to forecast the total electricity production in South Africa. It turns out that the comparison of the predictive performance of different forecasting methods is inevitable. Hybrid forecasting approaches, such as artificial neural network (ANN) based seasonal Autoregressive Integrated Moving Average (sARIMA) model, ANN based multiplicative Holt-Winters (HW) model, ANN based additive HW model, an adaptive neuro-fuzzy inference system (ANFIS) based sARIMA model, ANFIS based multiplicative HW model and ANFIS based additive HW model, are employed as some valuable alternatives compared with the conventional univariate time series models, such as sARIMA model and both multiplicative and additive HW models. The aim of this study is not only to provide evidence on the weakness of the univariate time series models, but also to show that hybrid forecasting method has the superior ability over the univariate time series models, with achieving a higher forecasting accuracy. In addition, random walk model is used as benchmark model, allowing for the fair competition. The results show that the hybrid model, ANN based on multiplicative HW model, is the most fitted for the total electricity production in South Africa. This study presents an empirical framework to guide the field of prediction research by providing a more comprehensive empirical investigation of the total electricity production forecasting by using various hybrid models.

    Citation: Emrah Gulay. Forecasting the total electricity production in South Africa: Comparative analysis to improve the predictive modelling accuracy[J]. AIMS Energy, 2019, 7(1): 88-110. doi: 10.3934/energy.2019.1.88

    Related Papers:

  • Electricity plays an important role in the South African economy with the industrial sector consuming the highest proportion followed by the residential and mining sector. Besides the fact that electricity is considered as an important energy sources, an adequate supply of electricity remains an important factor that affects the development and economic growth of a country. Therefore, it becomes even more important to forecast the total electricity production in South Africa. It turns out that the comparison of the predictive performance of different forecasting methods is inevitable. Hybrid forecasting approaches, such as artificial neural network (ANN) based seasonal Autoregressive Integrated Moving Average (sARIMA) model, ANN based multiplicative Holt-Winters (HW) model, ANN based additive HW model, an adaptive neuro-fuzzy inference system (ANFIS) based sARIMA model, ANFIS based multiplicative HW model and ANFIS based additive HW model, are employed as some valuable alternatives compared with the conventional univariate time series models, such as sARIMA model and both multiplicative and additive HW models. The aim of this study is not only to provide evidence on the weakness of the univariate time series models, but also to show that hybrid forecasting method has the superior ability over the univariate time series models, with achieving a higher forecasting accuracy. In addition, random walk model is used as benchmark model, allowing for the fair competition. The results show that the hybrid model, ANN based on multiplicative HW model, is the most fitted for the total electricity production in South Africa. This study presents an empirical framework to guide the field of prediction research by providing a more comprehensive empirical investigation of the total electricity production forecasting by using various hybrid models.


    加载中


    [1] Makridakis S, Hogarth RM, Gaba A (2010) Why forecasts fail. What to do instead. MIT Sloan Manage Rev 51: 83–90.
    [2] Khashei M, Bijari M (2010) An artificial neural network (p, d, q) model for timeseries forecasting. Expert Syst Appl 37: 479–489. doi: 10.1016/j.eswa.2009.05.044
    [3] Bianchi L, Jarrett J, Hanumara RC (1998) Improving forecasting for telemarketing centers by ARIMA modeling with intervention. Int J Forecast 14: 497–504. doi: 10.1016/S0169-2070(98)00037-5
    [4] De Gooijer JG, Hyndman RJ (2006) 25 years of time series forecasting. Int J Forecast 22: 443–473. doi: 10.1016/j.ijforecast.2006.01.001
    [5] Armstrong JS (2006) Findings from evidence-based forecasting: Methods for reducing forecast error. Int J Forecast 2: 583–598.
    [6] Chen CF, Chang YH, Chang YW (2009) Seasonal ARIMA forecasting of inbound air travel arrivals to Taiwan. Transportmetrica 5: 125–140. doi: 10.1080/18128600802591210
    [7] Gelper S, Fried R, Croux C (2010) Robust forecasting with exponential and Holt-Winters smoothing. J Forecast 29: 285–300.
    [8] Permanasari AE, Rambli DRA, Dominic PDD (2011) Performance of univariate forecasting on seasonal diseases: The case of tuberculosis. Software Tools Algorithm Biol Syst, 171–179.
    [9] Omane-Adjepong M, Oduro FT, Oduro SD (2013)Determining the better approach for short-term forecasting of Ghana's inflation: Seasonal ARIMA vs h. Int J Bus Humanities Technol 3: 69–79.
    [10] Rahman A, Ahmar AS (2017) Forecasting of primary energy consumption data in the United States: A comparison between ARIMA and Holter-Winters models. AIP Conf Proc 1885: 020163. doi: 10.1063/1.5002357
    [11] de Oliveira EM, Oliveira FLC (2018) Forecasting mid-long term electric energy consumption through bagging ARIMA and exponential smoothing methods. Energy 144: 776–788. doi: 10.1016/j.energy.2017.12.049
    [12] González JP, San Roque AM, Pérez EA (2018) Forecasting functional time series with a new Hilbertian ARMAX model: Application to electricity price forecasting. IEEE T Power Syst 33: 545–556. doi: 10.1109/TPWRS.2017.2700287
    [13] Lebotsa ME, Sigauke C, Bere A, et al. (2018) Short-term electricity demand forecasting using partially linear additive quantile regression with an application to the unit commitment problem. Appl Energ 222: 104–118. doi: 10.1016/j.apenergy.2018.03.155
    [14] Darbellay GA, Slama M (2000) Forecasting the short-term demand for electricity: Do neural networks stand a better chance? Int J Forecast 16: 71–83. doi: 10.1016/S0169-2070(99)00045-X
    [15] Chatfield C (1993) Neural networks: Forecasting breakthrough or passing fad? Int J Forecast 9: 1–3. doi: 10.1016/0169-2070(93)90043-M
    [16] Chatfield C (1995) Positive or negative? Int J Forecast 11: 501–502. doi: 10.1016/0169-2070(96)83105-0
    [17] Gorr WL, Nagin D, Szczypula J (1994) Comparative study of artificial neural network and statistical models for predicting student grade point averages. Int J Forecast 10: 17–34. doi: 10.1016/0169-2070(94)90046-9
    [18] Church KB, Curram SP (1996) Forecasting consumers' expenditure: A comparison between econometric and neural network models. Int J Forecast 12: 255–267. doi: 10.1016/0169-2070(95)00631-1
    [19] Callen JL, Kwan CC, Yip PC, et al. (1996) Neural network forecasting of quarterly accounting earnings. Int J Forecast 12: 475–482. doi: 10.1016/S0169-2070(96)00706-6
    [20] Tkacz G (2001) Neural network forecasting of Canadian GDP growth. Int J Forecast 17: 57–69. doi: 10.1016/S0169-2070(00)00063-7
    [21] Conejo AJ, Contreras J, Espinola R, et al. (2005) Forecasting electricity prices for a day-ahead pool-based electric energy market. Int J Forecast 21: 435–462. doi: 10.1016/j.ijforecast.2004.12.005
    [22] Chen T, Li L, Huang X (2005) Predicting the fibre diameter of melt blown nonwovens: Comparison of physical, statistical and artificial neural network models. Model Simul Mater Sc 13: 575. doi: 10.1088/0965-0393/13/4/008
    [23] Jain A, Kumar AM (2007) Hybrid neural network models for hydrologic time series forecasting. Appl Soft Comput 7: 585–592. doi: 10.1016/j.asoc.2006.03.002
    [24] Ding S, Hipel KW, Dang Y (2018) Forecasting China's electricity consumption using a new grey prediction model. Energy 149: 314–328. doi: 10.1016/j.energy.2018.01.169
    [25] Hu YC (2017) Electricity consumption prediction using a neural network based grey forecasting approach. J Oper Res Soc 68: 1259–1264. doi: 10.1057/s41274-016-0150-y
    [26] Tektaş M (2010)Weather forecasting using ANFIS and ARIMA models. Environ Res Eng Manage 51: 5–10.
    [27] Yayar R, Hekim M, Yilmaz V, et al. (2011) A comparison of ANFIS and ARIMA Techniques in the Forecasting of Electric Energy Consumption of Tokat Province in Turkey. J Econ Soc Stud 1: 87. doi: 10.14706/JECOSS11124
    [28] Yadav RK, Balakrishnan M (2014) Comparative evaluation of ARIMA and ANFIS for modeling of wireless network traffic time series. EURASIP J Wirel Commun Network 2014: 15. doi: 10.1186/1687-1499-2014-15
    [29] Hernandez CAS, Pedraza LFM, Salcedo OJP (2010) Comparative analysis of time series techniques ARIMA and ANFIS to forecast wimax traffic. Online J Electron Electric Eng 2: 223–228.
    [30] Lusis P, Khalilpour KR, Andrew L, et al. (2017) Short-term residential load forecasting: Impact of calendar effects and forecast granularity. Appl Energ 205: 654–669. doi: 10.1016/j.apenergy.2017.07.114
    [31] Luo J, Hong T, Fang SC (2018) Benchmarking robustness of load forecasting models under data integrity attacks. Int J Forecast 34: 89–104. doi: 10.1016/j.ijforecast.2017.08.004
    [32] Li C, Tao Y, Ao W, et al. (2018) Improving forecasting accuracy of daily enterprise electricity consumption using a random forest based on ensemble empirical mode decomposition. Energy 165: 1220–1227. doi: 10.1016/j.energy.2018.10.113
    [33] Chen Y, Kloft M, Yang Y, et al. (2018) Mixed kernel based extreme learning machine for electric load forecasting. Neurocomputing 312: 90–106. doi: 10.1016/j.neucom.2018.05.068
    [34] Zhang G, Patuwo BE, Hu MY (1998) Forecasting with artificial neural networks: The state of the art. Int J Forecast 14: 35–62. doi: 10.1016/S0169-2070(97)00044-7
    [35] Khashei M, Bijari M, Ardali GAR (2009) Improvement of auto-regressive integrated moving average models using fuzzy logic and artificial neural networks (ANNs). Neurocomputing 72: 956–967. doi: 10.1016/j.neucom.2008.04.017
    [36] Kunst RM (2012) Econometric forecasting. Institute for Advanced Studies Vienna and University of Vienna, Available from: http://homepage. univie. ac. at/robert. kunst/progpres. pdf.
    [37] Kler AM, Tyurina EA, Mednikov AS (2018) A plant for methanol and electricity production: Technical-economic analysis. Energy 165: 890–899. doi: 10.1016/j.energy.2018.09.179
    [38] World Meteorological Organization (2015) The Climate in Africa in 2013. WMO, No 1147.
    [39] Chellaney B (2013) Water, peace, and war: Confronting the global water crisis. Rowman & Littlefield.
    [40] Sparks D, Madhlopa A, Keen S, et al. (2014) Renewable energy choices and their water requirements in South Africa. J Energ South Afr 25: 80–92.
    [41] Holt CC (2004) Forecasting seasonals and trends by exponentially weighted moving averages. Int J Forecast 20: 5–13. doi: 10.1016/j.ijforecast.2003.09.015
    [42] Winters PR (1960) Forecasting sales by exponentially weighted moving averages. Manage Sci 6: 324–342. doi: 10.1287/mnsc.6.3.324
    [43] Box GE, Jenkins GM (1970) Time series analysis: Forecasting and control. San Francisco: Holden-Day.
    [44] Rumelhart DE, Hinton GE, Williams RJ (1986) Learning representations by back-propagating errors. Nature 323: 533. doi: 10.1038/323533a0
    [45] Jang JS (1993) ANFIS: Adaptive-network-based fuzzy inference system. IEEE T Syst Man Cy 23: 665–685. doi: 10.1109/21.256541
    [46] Makridakis S, Hibon M (1979) Accuracy of forecasting: An empirical investigation. J Roy Stat Soc 142: 97–125. doi: 10.2307/2345077
    [47] Rojas I, Valenzuela O, Rojas F, et al. (2008) Soft-computing techniques and ARMA model for time series prediction. Neurocomputing 71: 519–537. doi: 10.1016/j.neucom.2007.07.018
    [48] Pankratz A (1983) Forecasting with Univariate Box-Jenkins Models: Concepts and Cases. John Wiley and Sons, New York.
    [49] Makridakis S, Wheelwright SC, Hyndman RJ (2008) Forecasting methods and applications. John wiley & sons.
    [50] Kaboudan MA (2001) Compumetric forecasting of crude oil prices. Evolutionary Computation, 2001, Proceedings of the 2001 Congress on, IEEE, 1: 283–287. doi: 10.1109/CEC.2001.934402
    [51] Rasouli S, Tabesh H, Etminani K (2016) A Study of Input Variable Selection to Artificial Neural Network for Predicting Hospital Inpatient Flows. Brit J Appl Sci Techonol 18: 1–8.
    [52] Kecman V (2001) Learning and soft computing: support vector machines, neural networks, and fuzzy logic models. MIT press.
    [53] Kurkova V (1992) Kolmogorov's theorem and multilayer neural networks. Neural Networks 5: 501–506. doi: 10.1016/0893-6080(92)90012-8
    [54] Zhang P (2003) Time Series Forecasting using a hybrid ARIMA and neural network model. Neurocomputing 50: 159–175. doi: 10.1016/S0925-2312(01)00702-0
    [55] Diebold FX, Mariano RS (1995) Comparing Predictive Accuracy. J Bus Econ Stat 13: 253–263.
  • Reader Comments
  • © 2019 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Metrics

Article views(3934) PDF downloads(991) Cited by(3)

Article outline

Figures and Tables

Figures(12)  /  Tables(7)

Other Articles By Authors

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog