Abstract
This paper explores the application of the ARFIMA fractal model for prediction of
the dynamics of river water pollution based on BOD measure. The study begins by
conducting a review of related works in the field of water quality analysis. At this
stage also a suitable dataset is selected, that is used to train the ARFIMA, one of the
machine learning models. GPH semiparametric algorithm is applied for estimating the
fractal differentiation parameter of the ARFIMA. The obtained results are compared with
similar obtained with ARIMA model using RMSE and MAPE metrics. The study reveals an
enhancement in accuracy with the use of fractal methods for water pollution
prediction.
Publisher
Lviv Polytechnic National University
Reference10 articles.
1. [1] S. K. Jain, V. P. Singh, Water resources systems planning and management. Elsevier, 2023
2. [2] M. R. Penn, J. J. Pauer, and J. R. Mihelcic, "Biochemical oxygen demand." Environmental and ecological chemistry, vol. 2, 2009, pp. 278-297. ISBN: 978-1-84826-206-5. - P. 278-297.
3. [3] Y. Wen, G. Schoups, and N. van de Giesen, Organic pollution of rivers: Combined threats of urbanization, livestock farming and global climate change. Sci Rep 7, 43289 (2017). https://doi.org/10.1038/srep43289.
4. [4] Liu K, Chen Y, Zhang X. An Evaluation of ARFIMA (Autoregressive Fractional Integral Moving Average) Programs. Axioms. 2017; vol. 6(2), P. 1-16. https://doi.org/10.3390/axioms6020016
5. [5] B. K. Ray, Modeling long-memory processes for optimal long-range prediction, Journal of Time Series Analysis, 1993, vol. 14: pp. 511-525. https://doi.org/10.1111/j.1467-9892.1993.tb00161.x