中文摘要：

在基于网络行为分析的DDoS攻击检测中,人们往往把提升检测性能的研究重点放在提取更有效的属性特征、扩大检测范围、选用精度更高的机器学习算法等方面.文章则以DDoS攻击检测产生的报警信息为研究对象,通过一定的策略从大量的报警信息中去除误报警信息,从而达到降低误报的目的.

英文摘要：

In the DDoS attack detection based on network behavior analysis, people tend to improve the detection performance through extracting effective characteristics, expanding detection range, and selecting machine learning of higher accuracy. This paper studies the large amount of alarm information generated by DDoS attack detection and uses some strategies to remove the false alarm information, so as to reduce the false alarm ratio.