Random Sampling of Correlated Parameters – a Consistent Solution for Unfavourable Conditions

Journal Article · Thu Jan 15 00:00:00 EST 2015 · Nuclear Data Sheets

DOI:https://doi.org/10.1016/J.NDS.2014.12.032· OSTI ID:22436780

Trkov, A. ^[1]; Kodeli, I. A. ^[1]; Capote, R. ^[2]; Smith, D. L. ^[3]

---

1. Jožef Stefan Institute, Jamova cesta 39, SI-1000 Ljubljana (Slovenia)
2. International Atomic Energy Agency, PO Box 100, A-1400 Vienna (Austria)
3. Argonne National Laboratory, 1710 Avenida del Mundo, Coronado, CA 92118-3073 (United States)

Two methods for random sampling according to a multivariate lognormal distribution – the correlated sampling method and the method of transformation of correlation coefficients – are briefly presented. The methods are mathematically exact and enable consistent sampling of correlated inherently positive parameters with given information on the first two distribution moments. Furthermore, a weighted sampling method to accelerate the convergence of parameters with extremely large relative uncertainties is described. However, the method is efficient only for a limited number of correlated parameters.

• OSTI ID: 22436780
• Journal Information: Nuclear Data Sheets, Vol. 123; Conference: International workshop on nuclear data covariances, Santa Fe, NM (United States), 28 Apr - 1 May 2014; Other Information: Copyright (c) 2014 Elsevier Science B.V., Amsterdam, The Netherlands, All rights reserved.; Country of input: International Atomic Energy Agency (IAEA); ISSN 0090-3752
• Country of Publication: United States
• Language: English

Related Subjects

73 NUCLEAR PHYSICS AND RADIATION PHYSICS
DATA COVARIANCES
MULTIVARIATE ANALYSIS
NUCLEAR DATA COLLECTIONS
RANDOMNESS
STATISTICAL MODELS