CAJ | 학술논문

SVMlight算法和SMO算法在入侵检测领域的应用都是近年的研究热点．为了从中找到一种更适合入侵检测的算法，使用不同规模训练集和测试集进行多组实验，从不同角度研究它们在入侵检测中的特性．从检测精度、误报率和漏报率方面研究探讨两种算法各自在入侵检测中的优劣，并进行综合比较研究．找出了较优的算法为SMO算法．
SVMlight산법화SMO산법재입침검측영역적응용도시근년적연구열점．위료종중조도일충경괄합입침검측적산법，사용불동규모훈련집화측시집진행다조실험，종불동각도연구타문재입침검측중적특성．종검측정도、오보솔화루보솔방면연구탐토량충산법각자재입침검측중적우렬，병진행종합비교연구．조출료교우적산법위SMO산법．
The application of SVMlight algorithm and SMO algorithm in the field of intrusion detection is a research hotspot in recent years. Using different size training set and test set for several experiments, and study their features from different angles to find a more suitable algorithm for intrusion detection, and discuss their respective merits in intrusion detection from the detection accuracy, false positives and false negative rate, with a comprehensive comparative study to find out the optimum algorithm is the SMO algorithm.