Theory and Practice of Data Visualization for Modeling Complex Large Scale Data

该项目的目的是验证数据可视化技术的有效性，如Textile Plot，用于分析大型复杂数据，以发现潜在现象的新方面，并产生更好的建模。该项目是通过对真实的问题的真实的应用来进行的。该项目考虑了三个应用领域：基因组分析、金融数据分析和海洋调查分析。结果表明，与传统的LD显示方法相比，Textile Plot是发现新基因组的一个非常强大的工具。我们已经开发了一种新的算法，能够处理超高，例如几百万维的数据，这是常见的基因组分析。在金融领域，纺织品图是描述对冲基金操作差异和汇率差异的有力工具。我们与英联邦科学和工业研究组织在海洋调查数据分析领域开展了合作研究工作。将Textile Plot与PP Plot结合使用，可以发现环境变化对海底或河口动物影响的几个有趣方面。它使我们开发了一种新的分布模型的重量和数量的这种动物，作为一个平稳分布的随机差分方程的增长。

The aim of this project is to validate the effectiveness of data visualization technique like Textile Plot for analysis of large complex data to find out new aspects of underlying phenomena and result in a better modeling. The project has been conducted through real application to real problems. Three application fields are taken into account in this project, genome analysis, financial data analysis and marine survey analysis. It has been shown that Textile Plot is a quite powerful tool in finding new genomes compared with a traditional way of visualization LD Display. We have developed a new algorithm to be able to deal with super high, for example several million dimensional data, which is common in genome analysis. In the field of finance, it is shown that Textile Plot is a powerful tool in describing difference of hedge fund operations as well as FX rates. We have conducted cooperative research works with Commonwealth Science and Industry Research Organisation in the field of marine survey data analysis. We could find several interesting aspects of the effect of environmental changes to seabed or estuary fauna by using Textile Plot together with PP plots. It led us to develop a new distributional modeling of the weight and number of such animals, as a stationary distribution of a stochastic difference equation for the growth.