基于Jayaram Sethuraman在1994提出的扩展Dirichlet过程先验,将其推广到更一般的情形,使Dirichlet过程和扩展Dirichlet过程都成为一般Dirichlet过程的特例,并对非参数的贝叶斯进行了研究和讨论,给出了一般Dirichlet分布的期望、二阶矩等,证明了一般Dirichlet过程的支撑是足够大的,说明一般Dirichlet过程的构造是合理而又恰当的.
The paper gives a general class of nonparametric prior distributions on an arbitrary sample space.This paper gives the definition of general Dirichlet process,and proves that it is the prior distributions class in nonparametric problems.The research has discussed the properties of general Dirichlet process,and given the means /variances/covariances/finite moments of the prior class,then discussed the support of the prior class.