Hybrid Filter-Genetic Feature Selection Method For Arabic Sentiment Analysis
DOI:
https://doi.org/10.59167/tujnas.v8i1.1487Keywords:
Machin Learning, Sentiment Analysis, Opinion Mining, Feature Selection, ArabicAbstract
The dramatic increase in user comments describing their feelings about products, services, and events brings sentiment analysis to the forefront as a way to monitor public opinion about products and events. Feature selection is an important subtask of sentiment analysis, which aims to improve the performance of learning algorithms and reduce the dimensionality of a problem. Feature selection is an important subtask of sentiment analysis, as it can improve the performance of learning algorithms while reducing the dimensionality of a problem. Moreover, the high-dimensional feature spaces caused by the morphological richness of Arabic motivate further research in this area. In this paper, a hybrid filter-based and genetic feature selection algorithm is proposed using four machine learning algorithms, namely decision tree, Naive-Bayes, K-NN and meta-ensemble methods. The performance of the proposed algorithm is compared with the performance of baseline models. A wide range of experiments are conducted on two standard Arabic datasets. The experimental results clearly show that the improved methods outperform the other baseline models for Arabic sentiment analysis. The results show that the improved models outperform traditional approaches in terms of classification accuracy, with a 5% increase in the macro average of F1.
Downloads
Published
How to Cite
Issue
Section
Categories
License
Copyright (c) 2023 Muneer A.S. Hazaa, Saleh Ahmed Ali Hussein Salah
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
From July 2025 onward, all TUJNAS publications are licensed under the Creative Commons Attribution 4.0 International (CC BY 4.0) license. This open license allows anyone to share (copy and redistribute) and adapt (remix, transform, and build upon) the work in any medium or format, for any purpose (even commercially), as long as appropriate credit is given to the original author(s) and source. This permissive framework encourages scholarly innovation, translation, and integration into wider academic outputs by removing unnecessary legal barriers. Users of TUJNAS content must provide proper attribution and indicate if any changes were made to the original work. By enabling unrestricted reuse, the CC BY 4.0 license maximizes the reach and impact of research findings while ensuring that authors receive full recognition for their work. (For complete legal details of the CC BY 4.0 license, please refer to the official Creative Commons website.)
Submissions (from July 2025 onward): By submitting a manuscript to TUJNAS for publication (Volume 10, Issue 2, 2025 and thereafter), authors confirm the following:
- Originality: The submission is original, has not been published elsewhere, and is not under consideration by another journal.
- Copyright Retention: The author(s) retain copyright of the work, but grant TUJNAS a non-exclusive right to publish, reproduce, distribute, and archive the article.
- Open Access License: Upon acceptance, the article will be published open access under the CC BY 4.0 license.
- Repository Deposit: The author(s) agree that the full text and metadata of the article may be deposited in digital archives or repositories, to facilitate indexing and reuse under the CC BY 4.0 license.
- Indexing and Sharing: The author(s) acknowledge that TUJNAS may make the article available to third-party indexing, abstracting, and discovery services under the CC BY 4.0 license, without the need for additional permission.
These submission terms ensure that authors understand and consent to the open-access, licensed nature of TUJNAS publications from the outset.
