- 期刊
Research on Detection and Defense of Malicious Code Under Network Security
摘要
This paper mainly analyzed the detection and defense methods of malicious code, proposed to extract features by variable-length N-gram, and used weighted IG for feature selection. Finally, the performance of naive Bayesian (NB), random forest (RF), and support vector machine (SVM) classification models was compared. The results showed that the variable-length N-gram had a good performance in feature extraction, and weighted IG was better than IG in feature selection. Furthermore, the SVM model had the best performance in classifying malicious code and normal code, with an F1 score of 0.9367 and a log loss of 0.0321, which were better than the other two models. The results verify the reliability of the proposed method in the detection and defense of malicious code, which can be further applied in practice.
延伸閱讀
- 吳廸（2013）。A Study of Malicious Code Detection〔碩士論文，健行科技大學〕。華藝線上圖書館。https://www.airitilibrary.com/Article/Detail?DocID=U0022-0207201313565200
- Chang, Y. Y. (2010). Network Defense and Recovery Strategies for Maximization of Network Survivability under Malicious Attacks [master's thesis, National Taiwan University]. Airiti Library. https://doi.org/10.6342/NTU.2010.01731
- Kuo, C. T. (2018). A Security Risk Research of Software Defined Network Environment [doctoral dissertation, National Taiwan University]. Airiti Library. https://doi.org/10.6342/NTU201801596
- 梁德昭、莊李福（2010）。Study of Network Security Isolation technology-from the Perspective of Information Security Management。載於中華民國資訊安全學會（主編），全國資訊安全會議（頁240-245）。中華民國資訊安全學會。https://doi.org/10.29615/CISC.201005.0240
- 莊李福（2010）。Study of network security isolation technology - from the perspective of Information Security Management〔碩士論文，淡江大學〕。華藝線上圖書館。https://doi.org/10.6846/TKU.2010.00712