基于决策树算法的移动终端数据安全检测技术研究
2017-04-01刘江林袁宏彦
刘江林 袁宏彦



摘 要: 通过对决策树、k?Nearest Neighbor、贝叶斯三种不同数据挖掘算法的比较研究,基于可移动端数据的特点,建立了可移动端数据安全检测的模型框架,并通过实验对其加以验证。结果表明,决策树算法的检测分类结果最好,其查准率和查全率结果都很高;贝叶斯算法的检测分类结果性能稳定,但准确性不高,分类精度不理想,这是由该算法本身固有的特点决定的;k?Nearest Neighbor算法在开始时受到样本向量多少的影响,检测分类的效果不太稳定,分类效果在样本向量较少的情况下较差。通过对数据挖掘的可移动终端数据安全检测技术的研究,为今后数据安全检测技术的应用提供了一定的指导价值。
关键词: 数据挖掘; 移动终端; 数据安全; 检测技术
中图分类号: TN915.08?34 文献标识码: A 文章编号: 1004?373X(2017)05?0082?03
Abstract: By comparatively studying on the data mining algorithms of decision tree, k?Nearest Neighbor and Bayesian, a model framework of the mobile terminal data security detection was established according to the characteristics of the mobile terminal data, and verified with the experiment. The results show that the decision tree algorithm has the best detection and classification result, and its precision ratio and recall ratio are both high; the Bayesian algorithm has the stable performance of the detection and classification result, but its accuracy is low and classification precision is unsatisfied because of the inherent characteristics of the algorithm itself; the k?Nearest Neighbor algorithm reflected by the quantity of the sample vectors has unstable detection and classification result, and the classification result is poor when the algorithm has less sample vectors. The mobile terminal data security detection technology of the data mining is studied, which provides a certain guidance value for the application of the data security detection technology.
Keywords: data mining; mobile terminal; data security; detection technology
0 引 言
伴随着移动通信技术的飞速发展,移动终端在人们的日常生活中愈来愈多地承担互联网的应用和服务,但同时也带来了许多负面的影响,其中最大的挑战就是如何确保可移动端数据的安全[1?3]。可移动终端在承担以前PC端互联网的应用和服务时,自己也成了被攻击的对象,如何快速地检测、识别对可移动端数据存在安全威胁的数据,这一问题急需解决。
数据挖掘是将人工智能、机器学习、模式识别等多学科、多领域的知识结合,通过对当前大量信息数据的分析,找出各类事物之间新的联系和发展趋势等[4?7]。数据挖掘为解决可移动端数据安全监测问题提供了一种新的思路和途径,成为一个新的研究热点[8]。……
