Visualization and Virtual Materialization of Large Databases and Data Simulation based on 3D Meme Media

现在，科学技术的各个研究领域都在数据库中积累大量数据，使用最近开发的计算机控制的高效数据采集工具进行测量、分析和观测。研究人员认为，在数据库中积累如此庞大的数据将使他们能够在计算机上模拟各种物理、化学和/或生物现象，而不需要进行任何耗时和/或昂贵的真实实验。基于数据库的仿真的信息可视化要求每个可视化记录作为交互对象工作。当前的信息可视化系统将记录可视化，而不将其物化为交互对象。这些领域的研究人员根据他们的目标现象建立他们的个人或团体心理模型，并经常喜欢基于他们自己的心理模型来可视化信息。作为该项目的研究成果，我们提出了一个通用的开发数据库记录虚拟物化的框架，该框架基于我们于1995年开发的面向3D应用的组件体系结构SmarentBox。该框架提供了用于(1)访问数据库、(2)指定和修改数据库查询、(3)将交互式3D对象定义为模板以在虚拟空间中实现每个记录、以及(4)定义用于信息物化的虚拟空间及其坐标系的可视交互组件。这些组件被表示为框，即SmarentBox体系结构中的组件。我们还将该框架与统计数据挖掘和文本挖掘的研究成果相结合。

Various research fields in science and technology are now accumulating large amounts of data in databases, using recently developed computer controlled efficient data-acquisition tools for measurement, analysis, and observation. Researchers believe that such a huge extensive data accumulation in databases will allow them to simulate various physical, chemical, and/or biological phenomena on computers without carrying out any time-consuming and/or expensive real experiments. Information visualization for DB-based simulation requires each visualized record to work as an interactive object. Current information visualization systems visualize records without materializing them as interactive objects. Researchers in these fields develop their individual or community mental models on their target phenomena, and often like to visualize information based on their own mental models. As the research achievement of this program, we proposed a generic framework for developing virtual materialization of database records based on the component-ware architecture IntelligentBox that we developed in 1995 for 3D applications. This framework provides visual interactive components for (1) accessing databases, (2) specifying and modifying database queries, (3) defining an interactive 3D object as a template to materialize each record in a virtual pace, and (4) defining a virtual space and its coordinate system for the information materialization. These components are represented as boxes, i.e., components in the IntelligentBox architecture. We have also integrated this framework with statistical data mining and text mining research results.