Refining malware analysis with enhanced machine learning algorithms using hyperparameter tuning
Many researchers address challenges and limitations inherent to machine learning algorithms to optimize classifier performance. Overfitting, a prevalent issue, arises when models are excessively complex and trained on noisy data, leading to suboptimal generalization to new data. Another concern is u...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
InderScience
2024
|
| Subjects: | |
| Online Access: | https://repository.londonmet.ac.uk/8932/1/2023_IJCCBS-170512%20%285%29.pdf |
| _version_ | 1804072956240855040 |
|---|---|
| author | El Mouhtadi, Walid El Bakkali, Mohamed Maleh, Yassine Mounir, Soufyane Ouazzane, Karim |
| author_facet | El Mouhtadi, Walid El Bakkali, Mohamed Maleh, Yassine Mounir, Soufyane Ouazzane, Karim |
| author_sort | El Mouhtadi, Walid |
| collection | LMU |
| description | Many researchers address challenges and limitations inherent to machine learning algorithms to optimize classifier performance. Overfitting, a prevalent issue, arises when models are excessively complex and trained on noisy data, leading to suboptimal generalization to new data. Another concern is underfitting, where models are overly simplistic and fail to capture data complexity. This comprehensive investigation focuses on machine learning's application to malware classification, specifically targeting PE files. The study addresses these limitations using ensemble methods and pre-processing techniques, including feature selection and hyperparameter tuning. The primary objective is to augment classifier performance. Through a comparative study that aims to classify PE files as malicious or benign through analysis of machine learning methodologies such as random forests, decision trees, and gradient boosting, the study highlights the superiority of the random forests algorithm, achieving a remarkable 99% accuracy rate. Thoroughly assessing the strengths and limitations of each algorithm provides valuable insights into effectively handling diverse malware categories. This paper underscores the significance of ensemble methods, feature engineering, and pre-processing in enhancing classifier performance. |
| first_indexed | 2024-07-09T04:07:23Z |
| format | Article |
| id | oai:repository.londonmet.ac.uk:8932 |
| institution | London Metropolitan University |
| language | English |
| last_indexed | 2024-07-09T04:07:23Z |
| publishDate | 2024 |
| publisher | InderScience |
| record_format | eprints |
| spelling | oai:repository.londonmet.ac.uk:89322024-07-01T11:25:12Z http://repository.londonmet.ac.uk/8932/ Refining malware analysis with enhanced machine learning algorithms using hyperparameter tuning El Mouhtadi, Walid El Bakkali, Mohamed Maleh, Yassine Mounir, Soufyane Ouazzane, Karim 000 Computer science, information & general works 600 Technology Many researchers address challenges and limitations inherent to machine learning algorithms to optimize classifier performance. Overfitting, a prevalent issue, arises when models are excessively complex and trained on noisy data, leading to suboptimal generalization to new data. Another concern is underfitting, where models are overly simplistic and fail to capture data complexity. This comprehensive investigation focuses on machine learning's application to malware classification, specifically targeting PE files. The study addresses these limitations using ensemble methods and pre-processing techniques, including feature selection and hyperparameter tuning. The primary objective is to augment classifier performance. Through a comparative study that aims to classify PE files as malicious or benign through analysis of machine learning methodologies such as random forests, decision trees, and gradient boosting, the study highlights the superiority of the random forests algorithm, achieving a remarkable 99% accuracy rate. Thoroughly assessing the strengths and limitations of each algorithm provides valuable insights into effectively handling diverse malware categories. This paper underscores the significance of ensemble methods, feature engineering, and pre-processing in enhancing classifier performance. InderScience 2024-06-13 Article PeerReviewed text en https://repository.londonmet.ac.uk/8932/1/2023_IJCCBS-170512%20%285%29.pdf El Mouhtadi, Walid, El Bakkali, Mohamed, Maleh, Yassine, Mounir, Soufyane and Ouazzane, Karim (2024) Refining malware analysis with enhanced machine learning algorithms using hyperparameter tuning. International Journal of Critical Computer-Based Systems, 11 (1/2). pp. 48-67. ISSN 1757-8787 https://www.inderscienceonline.com/doi/10.1504/IJCCBS.2024.139100 10.1504/IJCCBS.2024.139100 |
| spellingShingle | 000 Computer science, information & general works 600 Technology El Mouhtadi, Walid El Bakkali, Mohamed Maleh, Yassine Mounir, Soufyane Ouazzane, Karim Refining malware analysis with enhanced machine learning algorithms using hyperparameter tuning |
| title | Refining malware analysis with enhanced machine learning algorithms using hyperparameter tuning |
| title_full | Refining malware analysis with enhanced machine learning algorithms using hyperparameter tuning |
| title_fullStr | Refining malware analysis with enhanced machine learning algorithms using hyperparameter tuning |
| title_full_unstemmed | Refining malware analysis with enhanced machine learning algorithms using hyperparameter tuning |
| title_short | Refining malware analysis with enhanced machine learning algorithms using hyperparameter tuning |
| title_sort | refining malware analysis with enhanced machine learning algorithms using hyperparameter tuning |
| topic | 000 Computer science, information & general works 600 Technology |
| url | https://repository.londonmet.ac.uk/8932/1/2023_IJCCBS-170512%20%285%29.pdf |
| work_keys_str_mv | AT elmouhtadiwalid refiningmalwareanalysiswithenhancedmachinelearningalgorithmsusinghyperparametertuning AT elbakkalimohamed refiningmalwareanalysiswithenhancedmachinelearningalgorithmsusinghyperparametertuning AT malehyassine refiningmalwareanalysiswithenhancedmachinelearningalgorithmsusinghyperparametertuning AT mounirsoufyane refiningmalwareanalysiswithenhancedmachinelearningalgorithmsusinghyperparametertuning AT ouazzanekarim refiningmalwareanalysiswithenhancedmachinelearningalgorithmsusinghyperparametertuning |