<<
>>

Exchangeable random sequences

Definition. We call two vectors x and y exchangeable if the empirical distri­bution of the two vectors are the same.

For more detail, see Zabell (1992).

2 These subscripts may be thought of as time indices or as giving the order in which samples are taken.

Note that the Polya urn model (Feller 1968, pp. 119-121) can produce the same conditional expectation. The mixing measure dμ in this case is the symmetrical Dirichlet prior with parameter α (Zabell 1992).

There is an obvious connection with the ways the entry and exit rates for agents are specified in our models. See the examples throughout this book (Chapters 4, 7, and 8 in particular) and Appendix A.3.

2.4

<< | >>
Source: Aoki M.. Modeling Aggregate Behaviour & Fluctuations in Economics. Cambridge: Cambridge University Press,2002. — 281 p.. 2002

More on the topic Exchangeable random sequences: