Topology-based Visual Analysis of Information Spaces

现代互联网技术为每个人提供了快速访问越来越多的文本文档，无论何时何地。这一发展对研究人员、企业和公众的信息处理提出了挑战。如果没有有效的手段让人类接触到大量的文献收藏，我们的社会就无法应对这一挑战，也无法利用这一机会来增加知识生产和加深理解。该项目将提供对大型动态电子文档集的高效、可扩展的访问。我们专注于主题演变的视觉分析，以获得内容的视觉概述以及交互式和直观的过滤机制。奋进需要自然语言处理和可视化研究的共同努力。我们将扩展我们成功的拓扑方法的第一个项目阶段，时间有序，不断扩大的文档集合。最终的用户界面将把文档表示为三维地形，其中山丘描述主题。

Modern internet technology provides fast access to an increasing amount of text documents for everyone, everywhere, at any time. This development challenges information processing of researchers, businesses, and general public. Without effective means for human access to large document collections, our society cannot meet this challenge and use the chance to increase knowledge production and create deeper understanding. This project will provide efficient, scalable access to large, dynamic electronic document collections. We focus on visual analysis of topic evolution to derive a visual overview of content together with interactive and intuitive filtering mechanisms. This endeavor requires a joint effort of natural language processing and visualization research. We will extend our successful topological approach of the first project phase to temporally ordered, expanding document collections. The final user interface will represent the documents as three-dimensional terrain in which hills describe topics.