一种层次结构中多维属性的可视化方法

发布时间：2018-06-27 01:49

本文选题：信息可视化 + 层次结构　；参考：《软件学报》2016年05期

【摘要】：在很多领域的统计分析中,通常需要分析既具有层次结构又具有多维属性的复杂数据,如食品安全数据、股票数据、网络安全数据等.针对现有多维数据和层次结构的可视化方法不能满足对同时具有层次和多维两种属性数据的可视分析要求,提出了一种树图中的多维坐标MCT(multi-coordinate in treemap)技术.该技术采用基于Squarified和Strip布局算法的树图表示层次结构,用树图中节点矩形的边作为属性轴,通过属性映射、属性点连接、曲线拟合实现层次结构中多维属性的可视化.将该技术应用于全国农药残留侦测数据,实现了对全国各地区、各超市、各农产品中农药残留检出和超标情况的可视化,为领域人员提供了有效的分析工具.MCT技术也可用于其他领域的层次多属性数据的可视化.
[Abstract]:In many fields of statistical analysis, we usually need to analyze complex data with hierarchical structure and multidimensional attributes, such as food safety data, stock data, network security data and so on. In view of the fact that the existing visualization methods of multidimensional data and hierarchical structure can not meet the requirements of visual analysis of data with both hierarchical and multidimensional attributes, a multi-dimensional coordinate MCT (multi-coordinate in treemap) technique in tree diagrams is proposed. Based on Squarified and Strip layout algorithm, tree graph is used to represent hierarchical structure, and the edge of node rectangle in tree map is used as attribute axis. Through attribute mapping, attribute point connection and curve fitting, the visualization of multidimensional attributes in hierarchical structure is realized. The technology is applied to the detection data of pesticide residue in the whole country, and the visualization of pesticide residue detection and exceeding standard is realized in all regions, supermarkets and agricultural products of the whole country. MCT can also be used to visualize hierarchical multi-attribute data in other fields.
【作者单位】：食品安全大数据技术北京市重点实验室(北京工商大学);北京工商大学计算机与信息工程学院;机器感知与智能教育部重点实验室(北京大学);北京大学信息科学与技术学院;Department
【基金】：“十二五”国家科技支撑计划(2012BAD29B01-2) 国家科技基础性工作专项(2015FY111200) 虚拟现实技术与系统国家重点实验室(北京航空航天大学)开放基金(BUAA-VR-14KF-04)~~
【分类号】：TP391.41

【参考文献】