Machine Learning and Cyber Threat Intelligence and Analytics
Permanent URI for this collection
1 - 4 of 4
ItemMalware Detection Using Frequency Domain-Based Image Visualization and Deep Learning( 2021-01-05)We propose a novel method to detect and visualize malware through image classification. The executable binaries are represented as grayscale images obtained from the count of N-grams (N=2) of bytes in the Discrete Cosine Transform (DCT) domain and a neural network is trained for malware detection. A shallow neural network is trained for classification, and its accuracy is compared with deep-network architectures such as ResNet that are trained using transfer learning. Neither dis-assembly nor behavioral analysis of malware is required for these methods. Motivated by the visual similarity of these images for different malware families, we compare our deep neural network models with standard image features like GIST descriptors to evaluate the performance. A joint feature measure is proposed to combine different features using error analysis to get an accurate ensemble model for improved classification performance. A new dataset called MaleX which contains around 1 million malware and benign Windows executable samples is created for large-scale malware detection and classification experiments. Experimental results are quite promising with 96% binary classification accuracy on MaleX. The proposed model is also able to generalize well on larger unseen malware samples and the results compare favorably with state-of-the-art static analysis-based malware detection algorithms.
ItemDeception Detection Using Machine Learning( 2021-01-05)Today’s digital society creates an environment potentially conducive to the exchange of deceptive information. The dissemination of misleading information can have severe consequences on society. This research investigates the possibility of using shared characteristics among reviews, news articles, and emails to detect deception in text-based communication using machine learning techniques. The experiment discussed in this paper examines the use of Bag of Words and Part of Speech tag features to detect deception on the aforementioned types of communication using Neural Networks, Support Vector Machine, Naïve Bayesian, Random Forest, Logistic Regression, and Decision Tree. The contribution of this paper is two-fold. First, it provides initial insight into the identification of text communication cues useful in detecting deception across different types of text-based communication. Second, it provides a foundation for future research involving the application of machine learning algorithms to detect deception on different types of text communication.
ItemA Meta-Model for Real-Time Fraud Detection in ERP Systems( 2021-01-05)Fraud is a worldwide issue affecting almost every organization once in a time. Recent studies have shown that fraudulent behavior impacts up to 5 % of a companies annual revenue. Information systems (IS) have become an integral part of every modern organization. They contain the data foundation of the entire company and thereby supporting business processes and day-to-day transactions. Although an IS usually contains control mechanisms to prevent different kinds of fraud, these mechanisms look insufficient, considering the role of IS in many fraud cases. Since many cases from different companies have shown the need for an appropriate countermeasure, we want to develop an application that efficiently detects fraud and fraudulent behavior. Therefore, we conducted a structured literature review and a qualitative survey to apply the design science research (DSR) methodology and derive requirements for a fraud detection system (FDS). As a result, we present a meta-model for a FDS for enterprise resource planning (ERP) systems. We also provide application requirements, principles, and features that define areas for further research.