Forecasting public expenditures using ARDL–LSTM hybrid model
DOI:
https://doi.org/10.62933/xr5r0296Keywords:
ARDL, LSTM, ARDL-LSTM, Public Revenues, Public Expenditures.Abstract
Public expenditure is regarded as the main tool for regulating fiscal policy and, consequently, affects the overall economic position of the country. This highlights the importance of high-quality forecasting, particularly when the objective is fiscal sustainability and the efficient allocation of resources. As economic behavior becomes more intricate and the relationships among variables increasingly intertwined, linear models are no longer sufficient for accurate prediction
With the assumption that public expenditure is conditioned by public revenues, this work will endeavor to predict government expenditure through the application of a hybrid ARDL-LSTM model. The method combines Autoregressive Distributed Lag (ARDL) model to explain linear variations and fluctuations of the time series and a neural network model-Long Short-Term Memory (LSTM) to explain nonlinear trends. Based on the comparison criteria. The findings indicated that the hybrid ARDL-LSTM model is more effective in forecasting accuracy, based on the RMSE and MAPE criterion using both the estimation and testing samples, which confirms that the combination of linear and nonlinear model is more efficient in predicting economic time series.
References
[1] M. Thenmozhi, “Forecasting stock index returns using ARIMA-SVM , ARIMA-ANN , and ARIMA-random forest hybrid models Manish Kumar *,” Int. J. Banking, Account. Financ., vol. 5, no. 3, pp. 284–308, 2014, doi: 10.1504/IJBAAF.2014.064307.
[2] F.-M. Tseng, G.-H. Tzeng, H.-C. Yu, and B. J. C. Yuan, “Fuzzy ARIMA model for forecasting the foreign exchange market,” Fuzzy Sets Syst., vol. 118, pp. 9–19, 2001, doi: https://doi.org/10.1016/S0165-0114(98)00286-3.
[3] M. L. De Prado, “Beyond Econometrics: A Roadmap Towards Financial Machine Learning,” SSRN Electron. J., pp. 1–32, 2020, doi: 10.2139/ssrn.3365282.
[4] M. H. Pesaran, Y. Shin, and R. J. Smith, “Bounds testing approaches to the analysis of level relationships,” J. Appl. Econom., vol. 16, pp. 289–326, 2001, doi: 10.1002/jae.616.
[5] N. Ozturk, A. Prof, K. Karagoz, and A. Prof, “Relationship Between Inflation and Financial Development: Evidence from Turkey Etkileri: Türkiye İçin Bulgular,” Int. J. Alanya Fac. Bus., vol. 4, no. 2, pp. 81–87, 2012, [Online]. Available: https://www.ajindex.com/dosyalar/makale/acarindex-1423869086.pdf
[6] M. E. P´erez-Pons, J. Parra-Dominguez, S. Omatu, E. Herrera-Viedma, and J. M. Corchado, “MACHINE LEARNING AND TRADITIONAL ECONOMETRIC MODELS : A SYSTEMATIC MAPPING STUDY,” J. Artif. Intell. Soft Comput. Res., vol. 12, no. 2, pp. 79–100, 2022, [Online]. Available: https://www.researchgate.net/publication/358879543_Machine_Learning_and_Traditional_Econometric_Models_A_Systematic_Mapping_Study
[7] D. Stempie and R. Slepaczuk, “Hybrid Models for Financial Forecasting : Combining Econometric , Machine Learning , and Deep Learning Models,” arXiv.org, pp. 1–30, 2022, [Online]. Available: https://arxiv.org/abs/2505.19617
[8] C. H. Aladag, E. Egrioglu, and C. Kadilar, “Improvement in Forecasting Accuracy Using the Hybrid Model of ARFIMA and Feed Forward Neural Network,” Am. J. Intell. Syst., vol. 2, no. 2, pp. 12–17, 2012, doi: 10.5923/j.ajis.20120202.02.
[9] M. A. Khan, H. Ali, H. Shabbir, F. Noor, and M. D. Majid, “Impact of Macroeconomic Indicators on Stock Market Predictions : A Cross Country Analysis,” J. Comput. Biomed. Informatics, vol. 08, no. 01, 2024, doi: doi.org/10.56979/801/2024.
[10] S. Hochreiter and J. Schmidhuber, “LONG SHORT-TERM MEMORY,” Neural Comput., vol. 9, no. 8, pp. 1–32, 1997, [Online]. Available: https://www.bioinf.jku.at/publications/older/2604.pdf
[11] F. A. Gers and N. N. Schraudolph, “Learning Precise Timing with LSTM Recurrent Networks,” J. Mach. Learn. Res., vol. 3, pp. 115–143, 2002, [Online]. Available: https://www.jmlr.org/papers/volume3/gers02a/gers02a.pdf
[12] K. Greff, R. K. Srivastava, J. Koutn, and B. R. Steunebrink, “LSTM : A Search Space Odyssey,” Trans. neural networks Learn. Syst., vol. 28, no. 10, pp. 2222–2232, 2016, doi: 10.1109/TNNLS.2016.2582924.
[13] Y. Yu, X. Si, C. Hu, and J. Zhang, “A Review of Recurrent Neural Networks : LSTM Cells and Network Architectures,” Neural Comput., vol. 31, pp. 1235–1270, 2019, doi: https://doi.org/10.1162/neco_a_01199.
[14] H. Abbasimehr and R. Paki, “Improving time series forecasting using LSTM and attention models,” J. Ambient Intell. Humaniz. Comput., no. 0123456789, 2021, doi: 10.1007/s12652-020-02761-x.
[15] S. M. Al-selwi, M. Fadzil, and S. Jadid, “Journal of King Saud University - Computer and Information Sciences RNN-LSTM : From applications to modeling techniques and beyond — Systematic review ✩,” J. King Saud Univ. - Comput. Inf. Sci., vol. 36, no. 5, p. 102068, 2024, doi: 10.1016/j.jksuci.2024.102068.
[16] X. Wei, L. Zhang, H. Yang, L. Zhang, and Y. Yao, “Geoscience Frontiers Machine learning for pore-water pressure time-series prediction : Application of recurrent neural networks,” Geosci. Front., no. January, 2020, doi: 10.1016/j.gsf.2020.04.011.
[17] M. Eren, “Fuzzy Autoregressive Distributed Lag model-based forecasting,” Fuzzy Sets Syst., vol. 459, pp. 82–94, 2023, doi: 10.1016/j.fss.2022.06.003.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Sulaiman Hussien Jawad, Munaf Yousif Hmood (Author)
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
Licensed under a CC-BY license: https://creativecommons.org/licenses/by-nc-sa/4.0/
