Topological Methods in Data Analysis and Visualization V - Theory, Algorithms, and Applications

Topological Methods in Data Analysis and Visualization V - Theory, Algorithms, and Applications
复制标题

数据分析和可视化中的拓扑方法 V - 理论、算法和应用

DOI:
10.1007/978-3-030-43036-8_7
复制
发表时间:
2020
期刊:
--
影响因子:
--
通讯作者:
Carr H
Carr H
中科院分区:
--
文献类型:
--
作者:
Carr H

文献摘要

相似文献

经过二十年的计算拓扑发展,这显然是一个计算上具有挑战性的领域。我们不仅在确定正确性方面遇到常见的算法和编程困难,而且还面临一类数学上复杂且符号上脆弱的问题。因此,有效的开发和部署需要一个额外的步骤——构建或选择合适的测试用例。由于我们无法测试所有可能的输入,因此我们对测试用例的选择表达了我们对任务和所涉及问题的理解。此外,我们使用的数据集的规模如此之大,无论这种行为在数学上多么不可能,它几乎肯定会在每次运行中大规模发生。因此,我们选择的测试用例与数学病态案例紧密结合,并且需要使用在构建数学反例时最明显表达的技能来开发。因此,本文首次尝试报告、分类和分析先前用于 Reeb 分析(等高线树和 Reeb 图)算法工作的测试用例,并表达如何测试拓扑代码的哲学。
After two decades of computational topology, it is clearly a computationally challenging area. Not only do we have the usual algorithmic and programming difficulties with establishing correctness, we also have a class of problems that are mathematically complex and notationally fragile. Effective development and deployment therefore requires an additional step—construction or selection of suitable test cases. Since we cannot test all possible inputs, our selection of test cases expresses our understanding of the task and of the problems involved. Moreover, the scale of the data sets we work with is such that, no matter how unlikely the behavior mathematically, it is nearly guaranteed to occur at scale in every run. The test cases we choose are therefore tightly coupled with mathematically pathological cases, and need to be developed using the skills expressed most obviously in constructing mathematical counter-examples. This paper is therefore a first attempt at reporting, classifying and analyzing test cases previously used for algorithmic work in Reeb analysis (contour trees and Reeb graphs), and the expression of a philosophy of how to test topological code.