Discrimination Analysis when the Variates are Grouped and Observed in Sequential Order
Abstract
Suppose that measurements x sub i = (x sub i sub 1, ...,X sub i sub j sub i), i = 1,...,k, can be taken on a unit sequentially in that order at the prescribed costs Ci, i = 1,...,k. The unit comes from one of the populations H sub 1 and H sub 2, and it is desired to select a population (from these two) from which the unit is supposed to belong to, on the basis of the measurements x sub 1, x sub 2, ... . Given the loss incurred by selecting population H sub i when in fact it belongs to H sub j, the prior probability p sub i of H sub i (i = 1,2), and assuming that H sub i has the normal distribution N (mu sub i, V), i = 1,2, the author derives the sequential Bayesian solution of the discrimination problem when mu sub 1, mu sub 2 and V are known. When mu sub i, V are unknown and must be estimated, He proposes a solution which is asymptotic Bayesian with exponential convergence rate. Keywords: Consistency, Exponential rate, Sequential procedure.
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 01, 1988
- Accession Number
- ADA198405
Entities
People
- Yuehua Wu
Organizations
- University of Pittsburgh