Fang, Lei
(2022)
Data Mining Approach to Understand Tensor Properties in Turbulent Cascade.
In: Pitt Momentum Fund 2022.
Abstract
Traditionally, turbulence study starts from unproven but plausible hypotheses (e.g. Kolmogorov's similarity hypotheses) and then proceeding in a systematic fashion to reach theory (e.g. K41). The results are then tested against the experimental data for correctness and completeness. While this hypothetical-deductive-like-method received many successes in history, it has its detrimental limitation: starting from a plausible hypothesis precludes any other possibilities. Since the turbulent data is the final test for turbulent theories, why not start by mining the turbulence data and seek their physical origin? The research objective is to use data mining approaches to gain a deeper understanding of the tensor properties in 2D turbulence. We choose 2D turbulence as the starting point of the data mining approaches for understanding turbulence because of the simpler geometries and the relative simplicity of the experimental data. Nevertheless, all the insights and machinery gained in 2D turbulence will be readily applicable to understand the 3D turbulence. Three major classes of data mining algorithms will be used are clustering algorithm, decision tree algorithm, and Apriori algorithm. We will tackle three well-defined questions in turbulence with the aforementioned algorithms.
Share
| Citation/Export: |
|
| Social Networking: |
|
Details
Metrics
Monthly Views for the past 3 years
Plum Analytics
Altmetric.com
Actions (login required)
 |
View Item |
![[feed]](/images/feed-icon-32x32.png){kind=icon}