 |
|
Related Communities:
|
 |
|
 |
Estimation of Scientific Hypotheses Quality in Virtual Experiments in Data Intensive Domains
| Author(s): | Evgeny Tarasov, Dmitry Kovalev |
| Created: | 2017/12/05 |
| Published: | Data Analytics and Management in Data Intensive Domains. Selected Papers of the XIX International Conference on Data Analytics and Management in Data Intensive Domains (DAMDID/RCDL 2017). CEUR Workshop Proceedings, ISSN 1613-0073, Vol. 2022. P. 281-292. |
| Abstract: | |
| In this paper, we investigate approaches that allow us to estimate the quality of model
implementing hypotheses within a virtual experiment. One of the areas of DID under study is the multiphase
fluid flow analyses. The quality estimation approach is carried out within the framework of the general
classical detection theory. The estimate is a binary indicator. As a instrument for testing hypotheses, two
approaches are used: frequency and Bayesian. Feature Extraction is carried out in the streaming mode. With
a certain periodicity, the quality assessment is recomputed taking into account the changing environmental
parameters. Thus, the moment when the model begins to poorly predict the behavior of the physical
phenomenon is captured. This paper describes the implementation of this approach within a distributed
computing framework. |
| Download: |
[ Adobe PDF ]
|
|
|
