Abstract: For the cognitive difficulties in network security when dealing with massive amounts of high-dimensional data, this paper proposes a method of visualization based on the improved right regular polygon of even sides. It takes the advantage of evaluation index for visual effect and correlation to rearrange dimension in order to optimize the display effect of dimension reduction. The experiments show that this method maintains the advantages of finding dimensional distribution information easily, and it overcomes the shortcomings of disorder of scatterplot and rare of presence information. It can help find cluster information effectively and facilitate data mining and analysis for administrator.

Key words: right regular polygon of even sides, visualization, dimension rearrangement, correlation coefficient, dimension reduction

摘要： 针对网络安全领域处理海量高维数据时遇到的认知困难，提出一种基于改进的正2k边形坐标的可视化方法，利用可视化效果评价指标和相关系数对维度进行重排，优化了高维数据在降维映射后的显示效果。实验结果表明，该方法保持了易于发现维度分布信息的优点，并解决了散点图数据混乱和呈现信息单一等缺点，能够更加有效地发现聚类信息，为管理员进一步分析和挖掘威胁提供了便利。

关键词: 正[2k]边形, 可视化, 维度重排, 相关系数, 降维映射