Undesirable effects of covariance matrix techniques for error analysis

Abstract

Regression with ${\chi }^2$ constructed from the covariance matrix should not be used for some combinations of covariance matrices and fitting functions. Using the technique for unsuitable combinations can amplify systematic errors. This amplification is uncontrolled, and can produce arbitrarily inaccurate results that might not be ruled out by a ${\chi }^2$ test. In addition, this technique can give incorrect (artificially small) errors for fit parameters. I give a test for this instability and a more robust (but computationally more intensive) method for fitting correlated data.

CERN-TH.6892/93 (revised), replaced because of major improvements (finding a test for stability of covariance matrix regression and an alternative method for fitting correlated data)

Cited by in corpus (9)

• On the Mass Spectrum of the SU(2) Higgs Model in 2+1 Dimensions
• Boundary Criticality of the 3D O($N$) Model: From Normal to Extraordinary
• Heavy Quark Spectroscopy and Matrix Elements: A Lattice Study using the Static Approximation
• Kardar-Parisi-Zhang universality class for the critical dynamics of reaction-diffusion fronts
• Fitting a function to time-dependent ensemble averaged data
• Spectral Density on the Lattice
• Computation of the dynamic critical exponent of the three-dimensional Heisenberg model
• Uncertainties due to imperfect knowledge of systematic effects: general considerations and approximate formulae
• Equilibrium and out-of-equilibrium critical dynamics of the three-dimensional Heisenberg model with random cubic anisotropy