U.S. Department of Energy

Pacific Northwest National Laboratory

Fishing for Features

In this project, researchers are developing a framework for discovering signature features of complex or poorly characterized phenomena. This involves a two-stage framework. In the first stage, a greedy strategy is used to discover new features that are potentially relevant to the underlying phenomenon (the source of such features being internet databases, scientific literature, government reports, weather, and scientific measurements, among others). This greedy strategy is likely to identify spurious correlations and produce false discovery. Therefore, in the second stage, features are pruned based on the strength of their association with the underlying model.

Project Staff: 
Kris Jarman, Alejandro Heredia-Langner, Brett Amidan, Tim Bays, Brian Clowers, John Cort, Bryan Linggi, Shari Matzner, Karen Wahl, David Wunschel
AttachmentSize
PDF icon SDI_Fishing_for_Features_flyer.pdf2.57 MB

Fishing for Features

The Fishing for Features Signature Discovery project developed a framework for discovering signature features in challenging environments involving large and complex data sets or where phenomena may be poorly characterized or understood. Researchers at PNNL have applied the framework to the optimization of biofuels blending and to discover signatures of climate change on microbial soil communities
| Pacific Northwest National Laboratory