Premium
Sequential sampling without recall from a dirichlet process
Author(s) -
Christensen Ronald
Publication year - 1989
Publication title -
canadian journal of statistics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.804
H-Index - 51
eISSN - 1708-945X
pISSN - 0319-5724
DOI - 10.2307/3315479
Subject(s) - reservation , recall , sampling (signal processing) , value (mathematics) , mathematics , context (archaeology) , observer (physics) , property (philosophy) , dirichlet distribution , computer science , mathematical optimization , statistics , psychology , cognitive psychology , philosophy , computer network , paleontology , physics , mathematical analysis , filter (signal processing) , epistemology , quantum mechanics , computer vision , biology , boundary value problem
Abstract The problem of sequential sampling without recall from a Dirichlet process is studied. An individual observes a sequential sample p 1 , p 2 , … . At each stage the observer can either accept the current value p i or continue sampling. The total cost to the observer is the current value plus a cost for each observation taken after the first. A concrete context for the problem is provided by a shopper who must buy a particular item and can elicit price quotations sequentially but must pay for each quotation. Qualitative properties of optimal search rules are derived which establish that the problem is well behaved. In particular, the reservation‐price property is shown to hold. The results extend those in Christensen (1986).