Assessment of Delivery Companies Based on Lexicon Using Intelligence Techniques and Arabic WordNet

Mohamad Ab.  Saleh; Ali Hadi  Hasan; Ahmed T.  Sadiq

doi:10.24996/ijs.2025.66.3.24

Authors

Mohamad Ab. Saleh Banking and Financial Department, College of Economic and Administration, Al Iraqia University, Baghdad, Iraq https://orcid.org/0000-0002-0061-0568
Ali Hadi Hasan College of Information Technology, University of Babylon, Babylon, Iraq
Ahmed T. Sadiq Department of Computer Science, University of Technology, Baghdad, Iraq

DOI:

https://doi.org/10.24996/ijs.2025.66.3.24

Keywords:

Arabic WordNet, Delivery companies, Lexicon-based, Sentiment Analysis, Machine Learning, NLP, Social media Data

Abstract

In recent years, the activities of delivery companies have surged due to various factors, including epidemics, cost-effectiveness, the ease of placing orders, and urban congestion. This paper presents a model for evaluating the performance of delivery companies by analyzing comments available on social media. Assessing a company's performance based on feedback from corporate customers provides a valuable indicator for predicting its future success or failure. The main strength of this paper lies in utilizing a simple dictionary of Arabic and Iraqi dialect words to describe essential terms and leveraging Arab WordNet to enhance the quality of description words extracted from comments. The accuracy values for the four machine learning models used in our study are as follows: Rough Set Theory (RST) 97.5%, Support Vector Machine (SVM) 96.7%, Random Forest (RF) 94.3%, and Naïve Bayes (NB) 89.2%. The results demonstrate that the RST algorithm outperforms the other algorithms in terms of accuracy, representation, and recall.