Abstract
This article aims to investigate the relationship between the available past and desired prediction spans of the compositional time series (CTS) data, while comparatively analyzing the performances of a classical and an alternative approach both based on Markov chain model, in the absence of clear literary guidelines. Contribution lies within the way of selection of optimal CTS proportion span of past and prediction data (past %/prediction %) rather than making a random choice. The various configurations of the experimental CTS data sets of energy proportions have been predicted using the two approaches followed by their mean absolute percentage error analyses within a common range of span proportions from 50–50% to 67–33% in a machine learning fashion. Prediction error has generally a decreasing tendency with increasing past data proportion; however, minimum error can lie anywhere within the tested range of span proportions. So, there is no ready-made choice of span proportion for the purpose of CTS prediction since there exists no clear empirical relationship between span proportion and prediction error. However, the current analysis suggests 59–41% and 58–42% span proportions along CTS as quick choices for comparable compositional prediction using classical and alternative approaches. The former approach is a better choice for CTS data prediction in spite of its inherent shortcoming. The rational may also be applied for the purpose of predictive modeling of other CTS data related to the applications, beyond energy sector, like agriculture, irrigation, marketing, supply chain, transportation, human resources and budget allocation, etc.
Similar content being viewed by others
References
Anonymous, "Statistical Review of World Energy 2020: Natural Gas," British Petroleum, London, UK, (2020)
Ediger, V.Ş; Akar, S.: ARIMA forecasting of primary energy demand by fuel in Turkey. Energy Policy 35(3), 1701–1708 (2007)
Li, W.; Lu, C.: The multiple effectiveness of state natural gas consumption constraint policies for achieving sustainable development targets in China. Appl. Energy 235, 685–698 (2019)
Anonymous, "State of Industry Reports 2009–2020," National electric power regulatory authority (NEPRA)-Government of Pakistan (GOP), Islamabad, 2009–2020.
Anonymous, "Pakistan Economic Surveys, 2014–2020 (statistical portion of annual reports)," Finance Division (FD), Government of Pakistan (GOP), Islamabad, 2014–2020.
Prek, M.: "Primary energy factor for electricity mix: the case of Slovenia," Energy Policy [Working Title], (2019)
Khan, M.A.: Modelling and forecasting the demand for natural gas in Pakistan. Renew. Sustain. Energy Rev. 49, 1145–1159 (2015)
Perwez, U.; Suhail, A.; Hasan, S.; Zia, U.: The long-term forecast of Pakistan’s electricity supply and demand: an application of long range energy alternatives planning. Energy 93, 2423–2435 (2015)
Synder, R.; Ord, J.; Kochler, K.A.B.; McLaren, K.R.; Beaumont, A.: Forecasting compositional time series: a state space. Int J Forecast 33, 502–512 (2017)
Benhaddi, M.; Ouarzazi, J.: Multivariate time series forecasting with dilated residual convolutional neural networks for urban air quality prediction. Arab J Sci Eng 46, 3423–3442 (2021)
Aitchison, J.: "The statistical analysis of compositional data. J Royal Stat Soci: Series B Methodol 44(2), 139–160 (1982)
Aitchison, J.; Egozcue, J.J.: Compositional data analysis: where are we and where should we heading? Math. Geol. 37, 829–850 (2005)
Sadiq, N.: Stochastic modelling of the daily rainfall frequency and amount. Arab J Sci Eng 39, 5691–5702 (2014)
Chen, Y.; Lin, X.; Zhang, S.: A proportion prediction model of terminal energy structure of IPS based on hidden markov Chain. Procedia CIRP 83, 456–460 (2019)
Ren, F.; Gu, L.: “Study on transition of primary energy structure and carbon emission reduction targets in china based on markov chain model and GM (1, 1),.” Math Probl Eng 2016(4912935), 8 (2016)
Ahmad, H.; Hayat, N.: Markov chain based modelling and prediction of natural gas allocation structure pattern in Pakistan. Int. J. Energy Sect. Manage. 14(5), 911–933 (2020)
Ahmad, H.; Hayat, N.: Modeling and prediction of primary energy supply and electricity generation structures based on Markov chain: an insight with focus on the role of natural gas in Pakistan. J. Chin. Inst. Eng. 44(2), 177–191 (2021)
Meidani, H.; Ghanem, R.: Multiscale markov models with random transitions for energy demand management. Energy Build 61, 267–274 (2013)
Ren, X.; Liu, Q.; Zhang, Y.: The proportion of energy consumption structure prediction based on markov chain. J Chem Pharmaceut Res (Online) 7(3), 378–382 (2015)
Billio, M.; Casarin, R.; Osuntuyi, A.: Markov switching GARCH models for bayesian hedging on energy futures markets. Energy Econ 70, 545–562 (2018)
Huang, W.-T.; Chen, G.; Chen, P.: Taiwan depository receipts forecasting along a novel regular markov chain model. J. Chin. Inst. Eng. 43(5), 458–466 (2020)
Wei, Y.; Wang, Z.; Wang, H.; Li, Y.: Compositional data techniques for forecasting dynamic change in China’s energy consumption structure by 2020 and 2030. J Clean Product (2020)
Kynčlová, P.; Filzmoser, P.; Hron, K.: Modeling compositional time series with vector autoregressive models. J. Forecast. 34(4), 303–314 (2015)
Koki, C.; Meligkotsidou, L.; Vrontos, I.: Forecasting under model uncertainty: Non-homogeneous hidden Markov models with Pòlya-Gamma data augmentation. J. Forecast. 39(4), 580–598 (2019)
Tian, R.; Shen, G.: Predictive power of Markovian models: evidence from US recession forecasting. J. Forecast. 38(6), 525–551 (2019)
Ruan, Z.; Luo, A.; Yao, H.: Stochastic time series analysis for energy system based on markov chain model. Mobile Net Appl 22(3), 427–434 (2017)
Li, Y.; He, Y.; Zhang, M.: Prediction of Chinese energy structure based on convolutional neural network-long short-term memory (CNN-LSTM). Energy Sci Eng 8, 2680–2689 (2020)
Kazemi, M.H.; Shiri, J.; Marti, P.; Majnooni-Heris, A.: Assessing temporal data partitioning scenarios for estimating reference evapotranspiration with machine learning techniques in arid regions. J Hydrol 590, 125252 (2020)
Wang, H.; Yao, Y.; Salhi, S.: Tension in big data using machine learning: analysis and applications. Technol. Forecast. Soc. Chang. 158, 160–175 (2020)
Liu, H.; Cocea, M.: Semi-random partitioning of data into training and test sets in granular computing contexts. Granular Comput 2, 357–386 (2017)
Beyca, O.; Ervural, B.; Tatuglo, E.; Ozuyar, P.; Zaim, S.: Using machine learning tools for forecasting natural gas consumption in the province of Istanbul. Energy Econ 80, 937–949 (2019)
He, Y.; Lin, B.: Forecasting china’s total energy demand and its structure using ADL-MIDAS Model. Energy 151, 420–429 (2018)
Acknowledgements
The authors would like to pay thanks to the authors of the published literature from which data sets have been referred and employed in this research study.
Funding
This research work is at the cost of the authors’ own income without obtaining any grant from the funding agencies within the public sector, private sector and not-for-profit organizations, etc.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The authors declare that this study did not involve any conflict of interests with anybody or any organization.
Human and Animal Rights
The authors also declare that this study did not involve any violation of human and animal rights.
Rights and permissions
About this article
Cite this article
Ahmad, H., Hayat, N. A Rationale for Past/Prediction Span Proportion in Markov Chain-Based Predictive Modeling of Energy-Related Compositional Time Series Data. Arab J Sci Eng 47, 15887–15898 (2022). https://doi.org/10.1007/s13369-022-06793-7
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s13369-022-06793-7