CAJ | 학술논문

现有的特征提取方法或不具有良好的抗噪性,或具有一定抗噪能力但特征提取的准确性较差.针对这一问题,利用多序列比算法能够求解序列间相同元素的全局最大一致性这一特点,提出一种包括多序列比对算法ClustalW+CSR、自适应噪声消除、特征转化这3个步骤的抗噪特征提取方法.实验结果表明,与当前其他基于网络的攻击特征自动提取方法相比较,该方法不仅特征提取的准确性较高,而且还具有良好的抗噪能力.
현유적특정제취방법혹불구유량호적항조성,혹구유일정항조능력단특정제취적준학성교차.침대저일문제,이용다서렬비산법능구구해서렬간상동원소적전국최대일치성저일특점,제출일충포괄다서렬비대산법ClustalW+CSR、자괄응조성소제、특정전화저3개보취적항조특정제취방법.실험결과표명,여당전기타기우망락적공격특정자동제취방법상비교,해방법불부특정제취적준학성교고,이차환구유량호적항조능력.
A automatic signature generation(ASG).Current ASG approaches suffer from the noises in the captured attack samples.To address this problem,a new noise-tolerant and accurate-signature generation approach was proposed,which incorporates the three steps:multiple sequence alignment algorithm ClustalW+CSR,adaptive noise elimination and signature transformation.Expedment results show that this approach not only can generate more accurate signatures for polymorphic network attacks compared with other methods,but also has better noise-tolerance.