Goodness-of-fit tests in many dimensions
arXiv:physics/0405008 · doi:10.1016/j.nima.2005.11.136
Abstract
A method is presented to construct goodness-of-fit statistics in many dimensions for which the distribution of all possible test results in the limit of an infinite number of data becomes Gaussian if also the number of dimensions becomes infinite. Furthermore, an explicit example is presented, for which this distribution as good as only depends on the expectation value and the variance of the statistic for any dimension larger than one.
14 pages
References in corpus (6)
- Comparison of different goodness-of-fit tests
- A new class of binning free, multivariate goodness-of-fit tests: the energy tests
- Gaussian limits for discrepancies. I: Asymptotic results
- Quantum field theory for discrepancies
- Quantum field theory for discrepancies II: 1/N corrections using fermions
- Loaded Dice in Monte Carlo: importance sampling in phase space integration and probability distributions for discrepancies