scientific article; zbMATH DE number 10337

From MaRDI portal
Publication:4713508





Let \(X_ 1,\ldots,X_ n\) be independent and identically distributed random variables with common pdf: \(p_ x=P(X_ 1=x)\), \(x\in\{1,\ldots,m\}\), where \(p_ 1+\ldots+p_ m=1\). A run in \(i\)'s of length \(t\hbox{ in }X=(X_ 1,\ldots,X_ n)\) is any sequence \((X_{j+1},\ldots,X_{j+1})\), where \(0\leq j\) and \(j+t\leq n\). Then, \(N_ k(i;k_ i)\) and \(M_ n(i;k_ i)\) are said to have type I and type II binomial distributions of order \(k\) with parameter \(p_ i\), respectively, if: (a) For each \(i\), \(N_ n(i;k_ i)\) is the number of runs in \(i\)'s of length \(k_ i\) in \(X\) counted in such a way that each \(X_ j=i\) can contribute at most one value to \(N_ k(i;k_ i)\). (b) For each \(i\), \(M_ n(i;k_ i)\) is the number of runs in \(i\)'s of length \(k_ i\) in \(X\), where all different runs in i's of length \(k_ i\) are counted. The authors introduce the joint distributions of \(N=(N_ n(1;k_ 1),\dots,N_ n(m;k_ m))\) and of \(M=(M_ n(1;k_ 1),\ldots,M_ n(m;k_ m))\) and they call them the \(m\)-variate type I and type II binomial distributions of order \(K=(k_ 1,\dots,k_ m)\) with parameters \(n\) and \(p=(p_ 1,\ldots,p_ m)\), respectively. They also establish recurrence relations among functions related to the probability distributions of two types of bivariate binomial distributions of order \(k\) and derive the covariance of the type II bivariate binomial distribution of order \(k\).











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 Q4713508)