The State of Cybersecurity in Digital Landscape to Unmask PDF Malware Attacks

One of the biggest challenges in cyber security today is detecting malware infections due to the growing sophistication of cyber threats. Using the CIC-PDFMal2022 dataset, this paper investigates the use of machine learning(ML) algorithms for the identification of malware assaults based on Portable Document Format (PDF) files. This dataset offers a wide range of attributes that capture many facets of PDF files and was specially selected for use in PDF malware investigation. This dataset is used to train ML algorithms such as Logistic Regression (LR), Decision Tree (DT), Adaptive Gradient Boosting (AdaBoost), and Extreme Gradient Boosting (XGBoost), which then use the patterns they identify as indicative of malicious behaviour, to accurately classify threats and take preventative measures against PDF-based cyberattacks. Examining attributes including file size, metadata properties, JavaScript presence, and other structural components specific to PDF files are all part of the research. By assessing how well ML models identify and mitigate PDF malware attacks, the study seeks to promote cybersecurity and increase the general adaptability of digital systems to changing cyber threats. Performance criteria like accuracy, recall, precision, and F1-score have been used to identify the most effective attack detection technique. With accuracy ratings of 98.58%, 98.99%, and 98.69%, respectively, the LR, AdaBoost, and DT models are in order of precedence, while the XGBoost model has the greatest accuracy rating of 99.84%.