On the comparison aspect of the Shannon entropy and the cogentropy

Author

Gueaieb, W. ; Karray, F. ; Kamel, M. ; Wong, A.K.C.

Author_Institution

Dept. of Syst. Design Eng., Waterloo Univ., Ont., Canada

Volume

2

fYear

1998

fDate

24-28 May 1998

Firstpage

621

Abstract

It is claimed that the way the cogentropy defines the probabilities of events, in a certain data set, assuming that only a possibly non-representative subset of the entire set is known, provides better estimates than using those conventionally applied in the Shannon (1948) entropy. In this paper, we apply both entropy measures to estimate the expected uncertainty in the data when the whole data set is known. This expected uncertainty value is nothing but the Shannon entropy when the entire set is known. The effect of the size of the subset on the approximations computed by both entropy measures is investigated. The study takes into consideration the conditional and unconditional cases

Keywords

entropy; image processing; probability; Shannon entropy; approximations; cogentropy; data set; entropy measures; event probabilities; expected uncertainty estimation; image; nonrepresentative subset; subset size; whole data set; Data analysis; Entropy; Equations; Measurement uncertainty; Probability; Random variables; Size measurement;

fLanguage

English

Publisher

ieee

Conference_Titel

Electrical and Computer Engineering, 1998. IEEE Canadian Conference on

Conference_Location

Waterloo, Ont.

ISSN

0840-7789

Print_ISBN

0-7803-4314-X

Type

conf

DOI

10.1109/CCECE.1998.685573

Filename

685573