scientific article; zbMATH DE number 4122932
The author considers a bidimensional random vector (X,Y) defined on a domain D having a finite or infinite measure. The probability density function (p.d.f.) of (X,Y) is f(x,y). First it is supposed that \(D=[a,b]\times [c,d]\) and \(\{x_ i\}_{1\leq i\leq n}\), \(\{y_ j\}_{1\leq j\leq m}\) are partitions of [a,b] and [c,d], respectively, where [a,b] and [c,d] are finite intervals and \(x_ i=a+(b-a)i/n\), \(1\leq i\leq n\), \(y_ j=c+(d-c)j/m\), \(1\leq j\leq m.\) The entropy \(H_ f^{(n,m)}\) as an ``approximation to the entropy \(H_ f\) of the random vector (X,Y) is defined, where \[ H_ f^{(n,m)}=-\sum^{n}_{i=1}\sum^{m}_{j=1}q_{ij}\log_ 2q_{ij}-l\quad og_ 2((nm)/(b-a)(d-c)), \] with \(q_{ij}=p_{ij}/P^{(n,m)}\), \(p_{ij}=f(u_ i,v_ j)(b-a)(d- c)/(nm)\), \(u_ i\in (x_{i-1},x_ i)\), \(v_ j\in (y_{j-1},y_ j)\), \(P^{(n,m)}=\sum^{n}_{i=1}\sum^{m}_{j=1}p_{ij}\). The author establishes a relation between \(H_ f^{(n,m)}\) and \(H_ f\). The case of \(measure(D)=\infty\) was also treated.
This page was built for publication:
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4735848)