Enhancing PDF Malware Detection through Logistic Model Trees

Computers, materials & continua/Computers, materials & continua (Print), Год журнала: 2024, Номер 78(3), С. 3645 - 3663

Опубликована: Янв. 1, 2024

Malware is an ever-present and dynamic threat to networks computer systems in cybersecurity, because of its complexity evasiveness, it challenging identify using traditional signature-based detection approaches.The study article discusses the growing danger cybersecurity that malware hidden PDF files poses, highlighting shortcomings conventional techniques difficulties presented by adversarial methodologies.The presents a new method improves virus document analysis Logistic Model Tree.Using dataset from Canadian Institute for Cybersecurity, comparative carried out with well-known machine learning models, such as Credal Decision Tree, Naïve Bayes, Average One Dependency Estimator, Locally Weighted Learning, Stochastic Gradient Descent.Beyond structural JavaScriptcentric analysis, research makes substantial contribution area boosting precision resilience detection.The use thorough feature selection approach, increased focus on file attributes all contribute efficiency paper emphasizes Tree's critical role tackling increasing threats proposes viable answer practical issues sector.The results reveal Tree superior, improved accuracy 97.46% when compared benchmark demonstrating usefulness addressing ever-changing landscape.

Язык: Английский

---

Examining the Performance of Various Pretrained Convolutional Neural Network Models in Malware Detection

Applied Sciences, Год журнала: 2024, Номер 14(6), С. 2614 - 2614

Опубликована: Март 20, 2024

A significant quantity of malware is created on purpose every day. Users smartphones and computer networks now mostly worry about malware. These days, detection a major concern in the cybersecurity area. Several factors can impact performance, such as inappropriate features classifiers, extensive domain knowledge, imbalanced data environments, computational complexity, resource usage. number existing methods have been impacted by these factors. Therefore, this paper, we will first identify determine best classifiers then use them order to propose method. The comparative strategy proposed procedure consist four basic steps: transformation (converting images from RGB grayscale), feature extraction (using ResNet-50, DenseNet-201, GoogLeNet, AlexNet, SqueezeNet models), selection PCA method), classification (including GDA, KNN, logistic, SVM, RF, ensemble learning), evaluation accuracy error metrics). Unbalanced Malimg datasets are used experiments validate efficacy results that were obtained. According comparison findings, KNN machine learning classifier. It outperformed other terms both error. In addition, DenseNet201 pretrained model dataset. DenseNet201-KNN had an rate 96% minimal 3.07%. surpass state-of-the-art approaches. computationally quicker than most since it uses lightweight design fewer vector dimensions.

Язык: Английский

---

A Survey on the Applications of Semi-supervised Learning to Cyber-security

ACM Computing Surveys, Год журнала: 2024, Номер 56(10), С. 1 - 41

Опубликована: Апрель 11, 2024

Machine Learning’s widespread application owes to its ability develop accurate and scalable models. In cyber-security, where labeled data is scarce, Semi-Supervised Learning (SSL) emerges as a potential solution. SSL excels at tasks challenging traditional supervised unsupervised algorithms by leveraging limited alongside abundant unlabeled data. This article presents comprehensive survey of in focusing on countering diverse cybercrimes, particularly intrusion detection. Despite potential, notable research gap persists, with few recent studies comprehensively reviewing SSL’s cyber-security. study examines state-of-the-art techniques tailored for cyber-security address this gap. Relevant methods are identified, their effectiveness evaluated empower researchers practitioners insights enhance measures. work sheds light addressing scarcity domains addition outlining new directions advance crucial field. By bridging gap, manuscript paves the way enhanced cyber-threat detection mitigation an increasingly interconnected world.

Язык: Английский

---

FACNN: fuzzy-based adaptive convolution neural network for classifying COVID-19 in noisy CXR images

Medical & Biological Engineering & Computing, Год журнала: 2024, Номер 62(9), С. 2893 - 2909

Опубликована: Май 7, 2024

Язык: Английский

---

A privacy-preserving federated learning with a secure collaborative for malware detection models using Internet of Things resources

Internet of Things, Год журнала: 2023, Номер 25, С. 101015 - 101015

Опубликована: Ноя. 30, 2023

Язык: Английский

---