Revenue's Forecasting of Aqaba Ports Company Using Wavelet Transform and ARIMA Models

  • S. AL wadi Department of Finance, Faculty of Business, the University of Jordan, Jordan
  • Osama Mohammad Al-Rawashdeh Department of Hotel Management & Tourism Science, Al-Balqa Applied University, Aqaba University College
  • Omar ALsinglawi Department of accounting, the University of Jordan/ Jordan,
  • Bara’ah A. ABU Dalwein Master degree student, Aqaba Company for Ports Operation & Management
  • Mohammad H. Saleh Department of Finance, Faculty of Business, the University of Jordan, Jordan
  • Jamil J. jaber Department of Finance, Faculty of Business, the University of Jordan, Jordan
  • Firas Al-Rawashdeh Department of Finance, Faculty of Business, the University of Jordan, Jordan

Keywords:

MODWT's Functions, Revenue, ARIMA Model, Forecasting ,

Abstract

This study aims to increase revenue forecasting accuracy by modeling a time series of monthly revenue data obtained from Aqaba Company for Ports Operations and Managements - Jordan from January 2011 to December 2020. Numerous mathematical functions are utilized in this investigation, including the non-linear spectral model, the maximal overlapping discrete wavelet transform (MODWT) based on the Coiflet function (C4), and the autoregressive integrated moving average (ARIMA). Significant findings of this work include an explanation of all previous events within the specified period and the development of a new forecasting model by merging the best MODWT function (C4) with the fitted ARIMA model. In addition, this study indicates how MODWT may dissect financial data, highlighting the most volatile events and enhancing the accuracy of forecasts. Root Mean Square Error (RMSE) will be utilized to assess forecasting ability. To make the empirical findings helpful to Aqaba Ports Company, we predict the results for the following five years. Modeling the data using the Maximum Overlapping Discrete Wavelet Transform (MODWT) based Coiflet function (C6) (ARIMA-MODWT (C6) with fitting ARIMA (2,1,2)). Therefore, all previous occurrences and fluctuations throughout the specified time frame will be described and explored in detail. MODWT-ARIMA (C6) is the best model compared to Haar, Daubechies (D4), Coiflet (C4), Least Symmetric (LA8), and Best Localized (Bl14) with fitted ARIMA model based on RMSE statistical criterion. This study highlights the deconstruction capability of MODWT by highlighting major events, fluctuations, and forecasts. The future values till the end of 2025 were then predicted.

Scroll to Top