本文選題：網(wǎng)絡(luò)安全態(tài)勢(shì) + 要素提取�。� 參考：《計(jì)算機(jī)應(yīng)用》2017年04期

【摘要】：網(wǎng)絡(luò)安全態(tài)勢(shì)要素選取的質(zhì)量對(duì)網(wǎng)絡(luò)安全態(tài)勢(shì)評(píng)估的準(zhǔn)確性起到至關(guān)重要的作用,而現(xiàn)有的網(wǎng)絡(luò)安全態(tài)勢(shì)要素提取方法大多依賴先驗(yàn)知識(shí),并不適用于處理網(wǎng)絡(luò)安全態(tài)勢(shì)數(shù)據(jù)。為提高網(wǎng)絡(luò)安全態(tài)勢(shì)要素提取的質(zhì)量與效率,提出一種基于屬性重要度矩陣的并行約簡算法,在經(jīng)典粗糙集基礎(chǔ)上引入并行約簡思想,在保證分類不受影響的情況下,將單個(gè)決策信息表擴(kuò)展到多個(gè),利用條件熵計(jì)算屬性重要度,根據(jù)約簡規(guī)則刪除冗余屬性,從而實(shí)現(xiàn)網(wǎng)絡(luò)安全態(tài)勢(shì)要素的高效提取。為驗(yàn)證算法的高效性,利用Weka軟件對(duì)數(shù)據(jù)進(jìn)行分類預(yù)測(cè),在NSL-KDD數(shù)據(jù)集中,相比利用全部屬性,通過該算法約簡后的屬性進(jìn)行分類建模的時(shí)間縮短了16.6%;對(duì)比評(píng)價(jià)指標(biāo)發(fā)現(xiàn),相比現(xiàn)有的三種態(tài)勢(shì)要素提取算法(遺傳算法(GA)、貪心式搜索算法(GSA)和基于條件熵的屬性約簡(ARCE)算法),該算法具有較高的召回率和較低的誤警率。實(shí)驗(yàn)結(jié)果表明,經(jīng)過該算法約簡的數(shù)據(jù)具有更好的分類性能,實(shí)現(xiàn)了網(wǎng)絡(luò)安全態(tài)勢(shì)要素的高效提取。
[Abstract]:The quality of selecting network security situation elements plays an important role in the accuracy of network security situation assessment. However, most of the existing network security situation elements extraction methods rely on prior knowledge and are not suitable for dealing with network security situation data.In order to improve the quality and efficiency of network security situation extraction, a parallel reduction algorithm based on attribute importance matrix is proposed. The idea of parallel reduction is introduced on the basis of classical rough set.The single decision information table is extended to several, and the attribute importance is calculated by using conditional entropy, and the redundant attributes are deleted according to the reduction rules, so that the network security situation elements can be extracted efficiently.In order to verify the efficiency of the algorithm, we use Weka software to classify and predict the data. In the NSL-KDD data set, compared with the use of all attributes, the reduced attributes of the algorithm shorten the time of classification modeling.Compared with the three existing situational element extraction algorithms (genetic algorithm, greedy search algorithm, GSAs) and conditional entropy based attribute reduction algorithm, this algorithm has higher recall rate and lower false alarm rate.The experimental results show that the data reduced by this algorithm has better classification performance and can efficiently extract the network security situation elements.
，

本文編號(hào)：1749221