Physics-based feature mining for large data exploration

Thompson, David S.; Machiraju, Raghu; Jiang, Ming; Nair, Jayashree; Craclun, G.; Subramanyam, A V

doi:10.1109/mcise.2002.1014977

Cited by 22 publications

(10 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our combined preliminary work on feature mining has been reported in the literature [7][8][9]. It should be noted that our work spans two application domains, namely CFD and molecular dynamics.…”

Section: Ongoing Workmentioning

confidence: 99%

See 1 more Smart Citation

Improvements to Response-Surface Based Vehicle Design Using a Feature-Centric Approach

Thompson

Parthasarathy

Machiraju

et al. 2004

Computational Science - ICCS 2004

View full text Add to dashboard Cite

Abstract. In this paper, we present our vision for a framework to facilitate computationally-based aerospace vehicle design by improving the quality of the response surfaces that can be developed for a given cost. The response surfaces are developed using computational fluid dynamics (CFD) techniques of varying fidelity. We propose to improve the quality of a given response surface by exploiting the relationships between the response surface and the flow features that evolve in response to changes in the design parameters. The underlying technology, generalized feature mining, is employed to locate and characterize features as well as provide explanations for feature-feature and feature-vehicle interactions. We briefly describe the components of our framework and outline two different strategies to improve the quality of a response surface. We also highlight ongoing efforts.

show abstract

Section: Ongoing Workmentioning

confidence: 99%

“…The second paradigm, based on an aggregate classification, is suitable for features with a more global influence such as vortices. More details can be found in [8][9].…”

Section: Ongoing Workmentioning

confidence: 99%

Improvements to Response-Surface Based Vehicle Design Using a Feature-Centric Approach

Thompson

Parthasarathy

Machiraju

et al. 2004

Computational Science - ICCS 2004

View full text Add to dashboard Cite

show abstract

“…A vortex is an example of a feature in fluid dynamics for which this approach is appropriate. Point classification techniques can also be used to locate vortices; however, without the verification step, false positives can pervade the analysis [40] and invariably cause erroneous conclusions to be drawn or meaningless models to be built.…”

Section: Aggregate Classification Paradigmmentioning

confidence: 99%

“…Denoising is often conducted to remove these features that may not be significant. This operation can be as simple as using a threshold for sizes of ROIs or can use a scalespace denoising technique [3,40]. The ROIs can then be ranked based on their size and an appropriate measure of feature strength (e.g., the density change for a shock or the total energy for a defect).…”

Section: Denoising Ranking and Trackingmentioning

confidence: 99%

See 1 more Smart Citation

Feature Mining Paradigms for Scientific Data

Jiang¹,

Choy²,

Mehta³

et al. 2003

Proceedings of the 2003 SIAM International Conference on Data Mining

View full text Add to dashboard Cite

Numerical simulation is replacing experimentation as a means to gain insight into complex physical phenomena. Analyzing the data produced by such simulations is extremely challenging, given the enormous sizes of the datasets involved. In order to make efficient progress, analyzing such data must advance from current techniques that only visualize static images of the data, to novel techniques that can mine, track, and visualize the important features in the data. In this paper, we present our research on a unified framework that addresses this critical challenge in two science domains: computational fluid dynamics and molecular dynamics. We offer a systematic approach to detect the significant features in both domains, characterize and track them, and formulate hypotheses with regard to their complex evolution. Our framework includes two paradigms for feature mining, and the choice of one over the other, for a given application, can be determined based on local or global influence of relevant features in the data.

show abstract