崔恒志 王紀軍 徐明生
關(guān)鍵詞: 移動作業(yè); 安全檢測; 決策樹; 數(shù)據(jù)分類; TF?IDF; 檢測率
中圖分類號: TN915.08?34 ? ? ? ? ? ? ? ? ? ? ? 文獻標識碼: A ? ? ? ? ? ? ? ? ? ? ? ?文章編號: 1004?373X(2019)03?0090?03
Abstract: The traditional security detection method can not effectively deal with the malicious intrusion problem of the mobile network. Therefore, a security detection algorithm based on ID3 decision tree algorithm is proposed. According to the operation model analysis of the mobile operating system, the corresponding safety detection model is designed. The weights of the sensitive words of abnormal content are calculated and sorted by means of TF?IDF, and the ID3 decision tree algorithm is used to classify the parsed data. The experimental results show that the proposed security detection algorithm is effective, and has higher detection rate than Naive Bayes algorithm.
Keywords: mobile operation; security detection; decision tree; data classification; TF?IDF; detection rate
作為企業(yè)日常生產(chǎn)管理的重要內(nèi)容,現(xiàn)場作業(yè)調(diào)度需要花費較多的人力和時間,尤其是人工操作完成的作業(yè)調(diào)度更是經(jīng)常發(fā)生錯誤,因此通過計算機輔助自動完成作業(yè)調(diào)度成為現(xiàn)在的主流,可以有效減少成本、提高生產(chǎn)效率。但是,隨著企業(yè)規(guī)模的日益擴大和移動網(wǎng)絡化的程度越來越高,移動網(wǎng)絡系統(tǒng)承載的業(yè)務也不斷增加,其安全問題也日益嚴峻[1?3]。不正當?shù)氖袌龈偁帉е潞诳蛺阂夤羝髽I(yè)移動作業(yè)系統(tǒng)的現(xiàn)象出現(xiàn),從而達到破壞企業(yè)正常生產(chǎn)的目的。
如何在保障移動作業(yè)系統(tǒng)正常運行的前提下,更好地實現(xiàn)入侵安全檢測和防護成為目前迫切需要解決的問題。現(xiàn)階段主要利用移動設備數(shù)據(jù)審計或者惡意程序檢測來確保移動終端系統(tǒng)的安全。但是,上述安全防護手段均存在較大局限性[4]。例如,利用移動設備數(shù)據(jù)審計的安全檢測通常局限于設備的IOS系統(tǒng)和品牌;惡意程序檢測也常常局限于固定類型的病系列,且必須實時更新病毒庫。
數(shù)據(jù)挖掘常用的算法包括ID3,Apriori,CN2等。隨著數(shù)據(jù)挖掘的廣泛應用,目前也出現(xiàn)了一些基于數(shù)據(jù)挖掘的檢測技術(shù)方法,如文獻[5]針對在云計算中DDoS攻擊的特點,設計出基于云計算的DDoS攻擊入侵檢測模型,將Apriori算法與K?means聚類算法相結(jié)合應用到入侵檢測模型中。文獻[6]對樸素貝葉斯算法進行改進,以此構(gòu)建入侵檢測數(shù)據(jù)挖掘模型,并運用該模型做入侵檢測,達到了80%以上的平均檢測準確率。但以上檢測方法均存在平臺兼容問題,且算法實現(xiàn)復雜度較高,運行計算開銷較大。
因此,本文提出一種基于ID3決策樹算法的安全檢測算法。上述不同數(shù)據(jù)挖掘安全入侵檢測算法,ID3決策樹算法具有結(jié)構(gòu)簡單、分類速度快且使用范圍廣等優(yōu)點,所以本文選擇其實現(xiàn)異常數(shù)據(jù)的分類。根據(jù)移動作業(yè)系統(tǒng)運行模型分析,設計了相應的安全檢測模型。通過TF?IDF對異常內(nèi)容的敏感詞進行權(quán)值計算和排序,并采用ID3決策樹算法對解析后的數(shù)據(jù)進行分類。實驗結(jié)果驗證了提出的安全檢測算法的有效性。


本文提出一種基于ID3決策樹算法的安全檢測算法。不同于傳統(tǒng)數(shù)據(jù)挖掘安全入侵檢測算法,ID3決策樹算法具有結(jié)構(gòu)簡單、分類速度快且使用范圍廣等優(yōu)點。通過TF?IDF對異常內(nèi)容的敏感詞進行權(quán)值計算和排序,實驗結(jié)果表明,相比于加權(quán)樸素貝葉斯算法,提出算法具有較高的檢測率和更低的誤報率,檢測率達到0.931,誤報率為0.053。
參考文獻
[1] PACINI E, MATEOS C, GARINO C G. Distributed job sche?duling based on swarm intelligence: a survey [J]. Computers & electrical engineering, 2014, 40(1): 252?269.
[2] NAVIMIPOUR N J, RAHMANI A M, NAVIN A H, et al. Job scheduling in the expert cloud based on genetic algorithms [J]. Kybernetes, 2014, 43(8): 1262?1275.
[3] WANG X, WANG Y, YUE C. A new multi?objective bi?level programming model for energy and locality aware multi?job scheduling in cloud computing [J]. Future generation computer systems, 2014, 36(7): 91?101.
[4] HANAMSAGAR A, BORATE B, JANE N, et al. Detection of firewall policy anomalies in real?time distributed network security appliances [J]. International journal of computer applications, 2015, 116(6): 215?221.
[5] 李博,宋廣軍.應用數(shù)據(jù)挖掘算法檢測云計算中的DDoS攻擊[J].齊齊哈爾大學學報(自然科學版),2014(6):1?4.
LI Bo, SONG Guangjun. Application of data mining algorithm to detect DDoS attacks in cloud computing [J]. Journal of Qiqihar University (natural science edition), 2014(6): 1?4.
[6] SEN S, DETECTION M, DETECTION A, et al. Using instance?weighted Naive Bayes for adapting concept drift in masquerade detection [J]. International journal of information security, 2014, 13(6): 583?590.
[7] NISHIMURA S. Optimal job scheduling of M/GI/1 queue with feedback: the discounted case [J]. Journal of the Operations Research Society of Japan, 2017, 31(3): 371?388.
[8] CHEN K, ZHANG Z, LONG J, et al. Turning from TF?IDF to TF?IGM for term weighting in text classification [J]. Expert systems with applications: an international journal, 2016, 66(C): 245?260.
[9] PHU V N, TRAN V T N, CHAU V T N, et al. A decision tree using ID3 algorithm for English semantic analysis [J]. International journal of speech technology, 2017, 20(4): 1?21.