Application of intelligent algorithms in library resource malicious download detection system

Xueqi Zhang
DOI: https://doi.org/10.3233/idt-240382
2024-06-01
Intelligent Decision Technologies
Abstract:The malicious download of library resources may lead to serious security risks and data leakage. To address this issue, this study proposes an intelligent algorithm based on Sliding Event Windows for detecting the malicious download behavior. This research method includes collecting and analyzing library resource download data and establishing a Sliding Event Window model for behavioral analysis. For each event window, this intelligent algorithm was utilized for feature extraction and behavior classification. Experimental results showed that window size and class radius had a large impact, while clustering radius had a small impact. The maximum topic cluster ratio affected the false alarm rate. In the ROC curve area comparison, the AUC values of the proposed method, RBM Method, a detection method based on IP request time interval, and a detection method based on IP request frequency were 0.904, 0.879, 0.841, and 0.797, respectively. The research confirmed that Sliding Event Windows can effectively improve the accuracy of malicious download detection for library resources, enhance the security of library resources, and protect user privacy. This study can further optimize resource utilization efficiency and promote scientific research and innovative development.
What problem does this paper attempt to address?