面向社会安全事件的分布式神经网络攻击行为分类方法

发布时间：2019-02-26 18:24

【摘要】：大数据时代下,社会安全事件呈现出数据多样化、数据量快速递增等特点,社会安全事件的事态与特性分析决策面临巨大的挑战。高效、准确识别社会安全事件中的攻击行为的类型,并为社会安全事件处置决策提供帮助,已经成为国家与网络空间安全领域的关键性问题。针对社会安全事件攻击行为分类,提出一种基于Spark平台的分布式神经网络分类算法(DNNC)。DNNC算法通过提取攻击行为类型的相关属性作为神经网络的输入数据,建立了各属性与攻击类型之间的函数关系并生成分布式神经网络分类模型。实验结果表明,所提出DNNC算法在全球恐怖主义数据库所提供的数据集上,虽然在部分攻击类型上准确率有所下降,但平均准确率比决策树算法提升15.90个百分点,比集成决策树算法提升8.60个百分点。
[Abstract]:In the era of big data, social security events show the characteristics of diversification of data and rapid increase of data quantity. The situation and characteristics of social security events are facing great challenges in analysis and decision-making. It has become a key issue in the field of national and cyberspace security to identify the types of attacks in social security incidents efficiently and accurately, and to provide assistance for the decision-making of social security incidents. Aiming at the attack behavior classification of social security events, a distributed neural network classification algorithm (DNNC) based on Spark platform is proposed. The DNNC algorithm extracts the related attributes of the attack behavior type as the input data of the neural network. The functional relationship between each attribute and attack type is established and a distributed neural network classification model is generated. The experimental results show that, although the accuracy of the proposed DNNC algorithm in the data set provided by the Global Terrorism Database has decreased in part of the attack types, the average accuracy is 15.90 percentage points higher than that of the decision Tree algorithm. It is 8.60 percentage points higher than the integrated decision tree algorithm.
【作者单位】：北京信息科技大学计算机学院;网络文化与数字传播北京市重点实验室;
【基金】：国家自然科学基金资助项目(61370065,61502040) 国家科技支撑计划项目(2015BAK12B00)~~
【分类号】：D035;TP183

【相似文献】