A correlative analysis process in a visual analytics environment

A correlative analysis process in a visual analytics environment
复制标题

可视化分析环境中的关联分析过程

DOI:
10.1109/vast.2012.6400491
复制
发表时间:
2012
期刊:
2012 IEEE Conference on Visual Analytics Science and Technology (VAST)
影响因子:
--
通讯作者:
Whitney K. Huang
Whitney K. Huang
中科院分区:
--
文献类型:
--
作者:
A. Malik;Ross Maciejewski;N. Elmqvist;Yun Jang;D. Ebert;Whitney K. Huang

文献摘要

被引文献

相似文献

在空间和时间数据集中发现模式和趋势一直是统计学和不同科学领域长期研究的问题。本文提出了一种可视化的分析方法,用于交互式探索和分析多变量数据集之间的时空相关性。我们的方法使用户能够发现相关性,并在数据集之间的不同时空聚合水平上探索潜在的因果或预测联系,并使他们能够了解分析之前的潜在统计基础。我们的技术利用皮尔逊的积矩相关系数和因素的领先或落后于不同的数据集之间检测的趋势和周期性模式。
Finding patterns and trends in spatial and temporal datasets has been a long studied problem in statistics and different domains of science. This paper presents a visual analytics approach for the interactive exploration and analysis of spatiotemporal correlations among multivariate datasets. Our approach enables users to discover correlations and explore potentially causal or predictive links at different spatiotemporal aggregation levels among the datasets, and allows them to understand the underlying statistical foundations that precede the analysis. Our technique utilizes the Pearson's product-moment correlation coefficient and factors in the lead or lag between different datasets to detect trends and periodic patterns amongst them.