移动云服务环境下的用户异常行为检测

发布时间：2018-01-05 19:19

本文关键词：移动云服务环境下的用户异常行为检测　出处：《河南科技大学》2017年硕士论文　论文类型：学位论文

【摘要】：移动云计算是随着云计算和移动互联网的不断发展和融合而产生的一种新型应用模式。这种新型应用模式突破了终端硬件的限制,使得移动终端更易实现便携数据存取、智能负载均衡,降低管理和按需服务成本。移动云服务带来便利的同时使得在“用户-环境-服务”三个层面更易出现安全问题,因此,如何在云服务进入实质性服务流程之前甄别用户身份及其行为的可信程度成为核心问题。为了解决上述问题,本文从用户可信性角度出发,采用递进方式研究移动云环境下的用户异常行为检测方法。本文的主要贡献与研究工作如下:1.针对识别异常用户行为常采用的聚类分析和相似性计算存在过拟合和特征信息淹没等问题,提出一种基于神经网络聚类的用户异常行为分析方法。该方法采用SVD分解进行降维和去噪,将信息熵引入神经网络的隐含层进行软化分,并利用信息熵求得的权重因子进行相似性计算。仿真结果表明,该方案可以明显提高检测速度和聚类精度,更适用于移动云环境。2.为了解决研究内容1中采样率不均造成的不平衡数据学习问题,提出一种基于信誉投票的用户行为异常协同分析方法。该方法结合异常检测与误用检测技术,利用采样率与剪枝技术构造训练样本,采用集成分类器进行投票以分类用户行为。在此基础上,基分类器根据信誉值对用户行为进行投票,投票机制遵循少数服从多数的原则,进一步提高识别速度的基础上提升了检测精度。3.为了能够提前判定用户行为的异常与否及用户意图,提出一种基于模式增长的异常行为识别与自主优化方法。该方法首先采用分层匹配的方法判定是否超出可信容忍范围;其次,采用功能流和数据流的分析方法对判定后的用户行为进行扩散分析;最后采用模式增长的方法构建较完备的正常结点子图集和异常影响结点子图集,实现对用户行为的自主提前判定。在移动云服务环境下识别异常行为已成为移动云计算领域的核心问题,本文针对用户的历史行为数据分析挖掘其行为模式,从而根据已有行为模式分析并识别异常行为。本文的研究方法保障了用户对服务实施的操作总是处于用户所属规则允许范围内,为移动云服务向用户提供低耗、高效、可靠的服务奠定了基础。
[Abstract]:Mobile cloud computing is a new application mode with the development and integration of cloud computing and mobile Internet, which breaks through the limitation of terminal hardware. It makes mobile terminal more easy to access portable data, intelligent load balancing. Reduce the cost of management and on-demand services. Mobile cloud services provide convenience while making it more vulnerable to security problems at the "user-environment-service" level. How to identify the user identity and the trustworthiness of the behavior before the cloud service enters the substantive service process becomes the core problem. In order to solve the above problems, this paper starts from the perspective of user credibility. A progressive approach is used to study the detection method of abnormal behavior of users in mobile cloud environment. The main contributions and research work of this paper are as follows:. 1. There are some problems in clustering analysis and similarity calculation, such as over-fitting and inundation of feature information, which are often used to identify abnormal user behavior. In this paper, a method of user abnormal behavior analysis based on neural network clustering is proposed, in which SVD decomposition is used to reduce and de-noising, and information entropy is introduced into the hidden layer of neural network for softening. The simulation results show that the proposed method can obviously improve the detection speed and clustering accuracy. It is more suitable for mobile cloud environment. 2. In order to solve the problem of unbalanced data learning caused by uneven sampling rate in research content 1. This paper presents a collaborative analysis method of user behavior anomaly based on reputation voting, which combines anomaly detection and misuse detection techniques, using sampling rate and pruning techniques to construct training samples. On the basis of this, the basic classifier votes the user behavior according to the credit value, and the voting mechanism follows the principle of the minority from majority. Further improve the speed of recognition on the basis of improved detection accuracy. 3. In order to be able to determine the user behavior in advance of abnormal or not and user intention. In this paper, a method of pattern growth based abnormal behavior recognition and autonomous optimization is proposed. Firstly, hierarchical matching is used to determine whether it is beyond the range of credibility and tolerance. Secondly, the analysis method of function flow and data stream is used to analyze the user behavior after judgment. Finally, the method of pattern growth is used to construct complete normal node subsets and abnormal impact node subsets. In the mobile cloud service environment, identifying abnormal behavior has become the core issue in the field of mobile cloud computing. This paper analyzes and mine the behavior patterns of users based on the historical behavior data. Thus according to the existing behavior pattern analysis and identification of abnormal behavior. The research method in this paper ensures that the user to implement the operation of the service is always within the scope of the user's rules, for mobile cloud services to provide users with low consumption. Efficient and reliable service laid the foundation.
【学位授予单位】：河南科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP393.09

【相似文献】