Premium
Discrete approximations to the information integral
Author(s) -
Ghurye S. G.,
Johnson B. R.
Publication year - 1981
Publication title -
canadian journal of statistics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.804
H-Index - 51
eISSN - 1708-945X
pISSN - 0319-5724
DOI - 10.2307/3315294
Subject(s) - discretization , mathematics , probability distribution , approximations of π , kullback–leibler divergence , line (geometry) , statistical physics , calculus (dental) , mathematical analysis , statistics , physics , geometry , medicine , dentistry
If f and g are probability densities on the line, the integral of f ln( f/g ) is the Kullback‐Leibler information of f with respect to g. In practice, the observations are discrete, and hence the actual distributions are discrete approximations to the theoretical continuous distributions. This paper is concerned with estimating loss of information due to discretization of data.