WCFG: A Weighted Control Flow Graph Dataset Design for Malware Classification

In cybersecurity, malware classification is essential for identifying and examining malicious software. This paper introduces the Weighted Control Flow Graph (WCFG) dataset to improve malware classification, which combines control flow graph analysis and signature-based variety. The dataset is created using a four-step methodology that combines the extraction of PE features, the detection of malicious functions, the creation of control flow graphs, and the merging of data with the application of weights. Labeled control flow graphs are presented as Cytoscape JSON files in the WCFG dataset, along with attributes like signature matches and weighting scores. The dataset allows researchers to compare different machine learning models, assess feature values, visualize control flow graphs, and investigate adversarial attacks by combining the strengths of signature-based detection and control flow graph analysis. The WCFG dataset helps to advance the field of malware classification and makes it easier to conduct additional cybersecurity research.