Analysis of data from continuous probability distributions

被引：20

作者：

Holy, TE ^{[1
]}

机构：

[1] PRINCETON UNIV,DEPT PHYS,PRINCETON,NJ 08544

来源：

PHYSICAL REVIEW LETTERS | 1997年 / 79卷 / 19期

关键词：

D O I：

10.1103/PhysRevLett.79.3545

中图分类号：

O4 [物理学];

学科分类号：

0702 ;

摘要：

Given a set of points drawn from an unknown continuous probability distribution, one often wishes to infer the underlying distribution. This distribution can be estimated using a simple scalar field theory. Fluctuations around the estimate are characterized by a robust measure of goodness of fit, analogous to the conventional chi(2), whose distribution can also be calculated. The resulting method of data analysis has some advantages over conventional approaches. [S0031-9007(97)04449-9].

引用

页码：3545 / 3548

页数：4

共 8 条

[1]

[Anonymous], ANN STAT

[2] Field theories for learning probability distributions [J].

Bialek, W ;

Callan, CG ;

Strong, SP .

PHYSICAL REVIEW LETTERS, 1996, 77 (23) :4693-4697

[3]

COLEMAN S, 1975, ASPECTS SYMMETRY, pCH7

[4]

Devroye L., 1987, A course in density estimation

[5]

Ghorai J., 1979, Journal of Statistical Computation and Simulation, V10, P65, DOI 10.1080/00949657908810347

[6] NONPARAMETRIC ROUGHNESS PENALTIES FOR PROBABILITY DENSITIES [J].

GOOD, IJ ;

GASKINS, RA .

BIOMETRIKA, 1971, 58 (02) :255-+

[7]

Shannon Claude E., 1964, The Mathematical Theory of Communication

[8] OPTIMIZED PERTURBATION-THEORY [J].

STEVENSON, PM .

PHYSICAL REVIEW D, 1981, 23 (12) :2916-2944

← 1 →