Abstract
The choice of a gain function in recursive estimation of prior probabilities using a mixture is discussed. In particular, the rate of convergence of two existing schemes is shown to be related to their respective gain functions. Numerical comparisons are made and an improved scheme is proposed.
Original language | English |
---|---|
Pages (from-to) | 497-498 |
Number of pages | 2 |
Journal | IEEE Transactions on Information Theory |
Volume | 26 |
Issue number | 4 |
DOIs | |
State | Published - Jul 1980 |
Externally published | Yes |
ASJC Scopus subject areas
- Information Systems
- Computer Science Applications
- Library and Information Sciences