多模态时空大数据可视分析方法综述

发布时间：2018-05-20 16:07

本文选题：时空大数据 + 可视分析　；参考：《测绘学报》2017年10期

【摘要】：时空大数据可视分析是近年国际大数据分析与数据可视化领域研究的热点前沿,也是全空间信息系统的核心研究内容之一。本文针对时空大数据多源、多粒度、多模态和时空复杂关联的特点,按照描述性、解释性和探索性3个层次分类归纳了现有时空大数据可视分析方法,论述了时空大数据可视分析中多模态特征筛选、新型人机交互分析以及探索性可视推理等技术难点和主要发展动态。研究表明,以数据可视化为主的描述性可视分析研究相对成熟,以可视环境下交互式挖掘分析实现问题诊断为主的解释性可视分析已成为当前大数据分析的焦点,而面向复杂问题协同决策的探索性可视分析方法则是大数据分析有待突破的重要发展方向。
[Abstract]:Spatiotemporal big data visualization analysis is a hot topic in the field of international big data analysis and data visualization in recent years, and it is also one of the core research contents of full spatial information system. In this paper, according to the characteristics of multisource, multi-granularity, multi-modal and spatio-temporal complex correlation of spatio-temporal big data, the existing spatio-temporal big data visual analysis methods are summarized according to three levels: descriptive, interpretive and exploratory. This paper discusses the technical difficulties and main development trends of multimodal feature selection, new human-computer interaction analysis and exploratory visual reasoning in spatio-temporal big data visual analysis. The research shows that the descriptive visual analysis based on data visualization is relatively mature, and the interpretive visual analysis based on interactive mining analysis to realize problem diagnosis in visual environment has become the focus of current big data analysis. The exploratory visual analysis method for collaborative decision making for complex problems is an important development direction of big data analysis.
【作者单位】：西南交通大学地球科学与环境工程学院;
【基金】：国家自然科学基金(41471320) 国家重点研发计划(2016YFB0502303)~~
【分类号】：P208

【相似文献】