Quantization for statistical classification
Abstract
A suitable expression for evaluating the Bhattacharyya distance for Markov chains was derived. Then, the quantization of the observed Markov sequence for sequential detection was examined. The criteria, used to construct the quantizer, were the minimizations of the average sample number under the hypothesis and the average sample number under the alternative. The problem of approximating the joint probability density function of the output n-dimensional vector of the quantizer was considered from a decision theory viewpoint.
- Publication:
-
Ph.D. Thesis
- Pub Date:
- September 1979
- Bibcode:
- 1979PhDT........73C
- Keywords:
-
- Digital Systems;
- Random Processes;
- Signal Processing;
- Analog To Digital Converters;
- Markov Chains;
- Statistical Analysis;
- Communications and Radar