Goodness-of-fit criteria for the Adams and Jefferson rounding methods and their limiting laws (Q851233)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Goodness-of-fit criteria for the Adams and Jefferson rounding methods and their limiting laws |
scientific article |
Statements
Goodness-of-fit criteria for the Adams and Jefferson rounding methods and their limiting laws (English)
0 references
17 November 2006
0 references
Let \(W = (w_1, \dots, w_c)\) be some arbitrary probability vector for a fixed number of categories \(c\). The arbitrariness is modelled by assuming \(W\) to be random and, moreover, to follow a continuous distribution on the probability simplex \[ S_c = \{(w_1, \dots, w_c) \in [0, 1]^c : w_1 + \dots + w_c = 1\}. \] Rounding methods are used to round the continuous weights \(W_i\) to rational proportions of the form \(N_i/n\) for some prescribed integer accuracy \(n\). Implementing the classical methods of Adams and Jefferson, the authors investigate goodness-of-fit criteria for this rounding process. Assuming that the given probabilities are uniformly distributed, the limiting laws of the criteria are derived, first when the rounding accuracy increases, and then when the number of categories grows large.
0 references
appointment method
0 references
rounding error analysis
0 references
Gaussian limit law
0 references
q-stationary multiplier method
0 references
Sainte-Laguë divergence
0 references
convergence in distribution
0 references
goodness-of-fit criteria
0 references
0 references