Speaker
Nick Smith
(Fermi National Accelerator Lab. (US))
Description
For binned likelihood fits, a (dimension-6) quadratic EFT parameterization of the bin yield can be represented as a matrix norm. Often, this matrix is low-rank, presenting a possibility for more efficient computation in large fits. Analyzing the eigenspectrum of this matrix may also provide a path towards incremental optimization of binning, providing an intermediate approach to EFT analysis design between STXS-like and optimal observable-based approaches.
Author
Nick Smith
(Fermi National Accelerator Lab. (US))
