2022_An Enhanced Grasshopper Optimization Algorithm With Neural Networks For Improving The Accuracy Of Spam Email Detection
| Format: | General Document |
|---|
| _version_ | 1860798147101458432 |
|---|---|
| building | INTELEK Repository |
| collection | Online Access |
| collectionurl | https://intelek.unisza.edu.my/intelek/pages/search.php?search=!collection3 |
| copyright | Copyright©PWB2025 |
| country | Malaysia |
| date | 2022-08-08 |
| format | General Document |
| id | 16175 |
| institution | UniSZA |
| originalfilename | 16175_684606870bfd75d.pdf |
| person | Sanaa Abduljabbar Ahmed Ghaleb |
| recordtype | oai_dc |
| resourceurl | https://intelek.unisza.edu.my/intelek/pages/view.php?ref=16175 |
| sourcemedia | Server storage Scanned document |
| spelling | 16175 https://intelek.unisza.edu.my/intelek/pages/view.php?ref=16175 https://intelek.unisza.edu.my/intelek/pages/search.php?search=!collection3 General Document Malaysia Library Staff (Top Management) Library Staff (Management) Library Staff (Support) Terengganu Faculty of Informatics & Computing English application/pdf 1.5 Server storage Scanned document Universiti Sultan Zainal Abidin UniSZA Private Access UNIVERSITI SULTAN ZAINAL ABIDIN SAMBox 2.3.4; modified using iTextSharp™ 5.5.10 ©2000-2016 iText Group NV (AGPL-version) Copyright©PWB2025 2022-08-08 334 16175_684606870bfd75d.pdf Sanaa Abduljabbar Ahmed Ghaleb Spam (Electronic mail) Spam Email Detection Grasshopper Optimization Algorithm (GOA) Neural Networks 2022_An Enhanced Grasshopper Optimization Algorithm With Neural Networks For Improving The Accuracy Of Spam Email Detection Spam email is the unsolicited messages sent in large quantities over the Internet. A significant negative impact of spam email is not only limited to the waste of resources, time, and effort, but also increases communications overload and cybercrime. Spam detection is highly needed to prevent negative usages to subjected emails recently. Hence the need for better spam detection model to achieve better spam detection accuracy is relevant. Previous research shows that although techniques such as neural network, support vector machine and k-nearest neighbor classification have demonstrated the ability to detect spam, their performance and optimization in detecting spam still need more work in improving accuracy. Neural Networks (NNs) are one of the most popular techniques to perform non-linear classification and have been extensively used in the literature to perform spam detection. However, the training datasets usually compose feature sets of irrelevant or redundant information, which impacts the performance of classification, and traditional learning algorithms such as backpropagation suffer from known issues, including slow convergence and the trap of local minimum. Those problems lend themselves to the realm of optimization. Considering the wide success of swarm intelligence methods in optimization problems, the main objective of this thesis is to contribute to the improvement of spam detection technology through the application of swarm-based optimization techniques to the basic problems of selecting optimal packet features, and optimal training of neural networks on classifying those features into normal and attack instances. Therefore, this study proposes an enhanced spam email detection model by adapting enhanced Grasshopper Optimization Algorithm (GOA) with neural networks and feature selector. The research methodology encompassed three main stages. The first stage is to formulate suitable algorithm to be adapted in neural network training for the purpose of spam detection. In this stage, the GOA has been modified into six different variants of the mutation operator that enhances the diversity of the standard GOA in order to optimize neural network parameters. In the second stage, all algorithms are trained using multilayer perceptron neural networks extensively. The results of training performance are then evaluated and compared with current and traditional metaheuristics algorithms. The final stage is the classification task which is based on a new multi-objective binary GOA algorithm (MOBGOA) for wrapper-approach-based feature selection. This classification task is accomplished by selecting the optimal subset features of spam and tested with all succeeding algorithms from previous stage. These models are tested and applied to SpamBase, SpamAssassin, and UK-2011 Webspam benchmark datasets. The result is analysed and compared with similar studies in the literature for the detection of spam email. The result shows that the performance of the proposed MOBGOA algorithm achieved 97.5%, 98.3%, and 96.4% of accuracy scores compared to other techniques for the respective datasets. All developed models recorded false alarm rates optimized under value 0.1 which are 0.033, 0.018, and 0.043; and detection rate scores of 98.1%, 98.3% and 97.2% for the respective datasets. These results also show comparable results with previous research, which achieved a maximum of 96% accuracy. The findings represent the evidence that the proposed enhanced algorithm and the spam detection model achieved better level of optimization in spam detection than the state of the art. The developed and extended approach can robustly be implemented in detecting other malicious attacks such as phishing, cross-site scripting, malware, and botnets. Dissertations, Academic Thesis |
| spellingShingle | 2022_An Enhanced Grasshopper Optimization Algorithm With Neural Networks For Improving The Accuracy Of Spam Email Detection |
| state | Terengganu |
| subject | Spam (Electronic mail) Dissertations, Academic |
| summary | Spam email is the unsolicited messages sent in large quantities over the Internet. A significant negative impact of spam email is not only limited to the waste of resources, time, and effort, but also increases communications overload and cybercrime. Spam detection is highly needed to prevent negative usages to subjected emails recently. Hence the need for better spam detection model to achieve better spam detection accuracy is relevant. Previous research shows that although techniques such as neural network, support vector machine and k-nearest neighbor classification have demonstrated the ability to detect spam, their performance and optimization in detecting spam still need more work in improving accuracy. Neural Networks (NNs) are one of the most popular techniques to perform non-linear classification and have been extensively used in the literature to perform spam detection. However, the training datasets usually compose feature sets of irrelevant or redundant information, which impacts the performance of classification, and traditional learning algorithms such as backpropagation suffer from known issues, including slow convergence and the trap of local minimum. Those problems lend themselves to the realm of optimization. Considering the wide success of swarm intelligence methods in optimization problems, the main objective of this thesis is to contribute to the improvement of spam detection technology through the application of swarm-based optimization techniques to the basic problems of selecting optimal packet features, and optimal training of neural networks on classifying those features into normal and attack instances. Therefore, this study proposes an enhanced spam email detection model by adapting enhanced Grasshopper Optimization Algorithm (GOA) with neural networks and feature selector. The research methodology encompassed three main stages. The first stage is to formulate suitable algorithm to be adapted in neural network training for the purpose of spam detection. In this stage, the GOA has been modified into six different variants of the mutation operator that enhances the diversity of the standard GOA in order to optimize neural network parameters. In the second stage, all algorithms are trained using multilayer perceptron neural networks extensively. The results of training performance are then evaluated and compared with current and traditional metaheuristics algorithms. The final stage is the classification task which is based on a new multi-objective binary GOA algorithm (MOBGOA) for wrapper-approach-based feature selection. This classification task is accomplished by selecting the optimal subset features of spam and tested with all succeeding algorithms from previous stage. These models are tested and applied to SpamBase, SpamAssassin, and UK-2011 Webspam benchmark datasets. The result is analysed and compared with similar studies in the literature for the detection of spam email. The result shows that the performance of the proposed MOBGOA algorithm achieved 97.5%, 98.3%, and 96.4% of accuracy scores compared to other techniques for the respective datasets. All developed models recorded false alarm rates optimized under value 0.1 which are 0.033, 0.018, and 0.043; and detection rate scores of 98.1%, 98.3% and 97.2% for the respective datasets. These results also show comparable results with previous research, which achieved a maximum of 96% accuracy. The findings represent the evidence that the proposed enhanced algorithm and the spam detection model achieved better level of optimization in spam detection than the state of the art. The developed and extended approach can robustly be implemented in detecting other malicious attacks such as phishing, cross-site scripting, malware, and botnets. |
| title | 2022_An Enhanced Grasshopper Optimization Algorithm With Neural Networks For Improving The Accuracy Of Spam Email Detection |
| title_full | 2022_An Enhanced Grasshopper Optimization Algorithm With Neural Networks For Improving The Accuracy Of Spam Email Detection |
| title_fullStr | 2022_An Enhanced Grasshopper Optimization Algorithm With Neural Networks For Improving The Accuracy Of Spam Email Detection |
| title_full_unstemmed | 2022_An Enhanced Grasshopper Optimization Algorithm With Neural Networks For Improving The Accuracy Of Spam Email Detection |
| title_short | 2022_An Enhanced Grasshopper Optimization Algorithm With Neural Networks For Improving The Accuracy Of Spam Email Detection |
| title_sort | 2022_an enhanced grasshopper optimization algorithm with neural networks for improving the accuracy of spam email detection |