Title :
Histogram classification using vector quantization
Author :
Nobel, Andrew B. ; Lugosi, Gábor
Author_Institution :
Beckman Inst. for Adv. Sci. & Technol., Illinois Univ., Urbana, IL, USA
fDate :
27 Jun-1 Jul 1994
Abstract :
Examines a general scheme by which vector quantizer (VQ) design methods can be used to produce a histogram classification rule. Results on histogram classification using data-dependent partitions show that a number of VQ common design methods yield classification rules that are Bayes risk consistent. A new design method that accounts of misclassification errors is also considered
Keywords :
Bayes methods; vector quantisation; Bayes risk; data-dependent partitions; histogram classification; histogram classification rule; misclassification errors; vector quantization; Artificial intelligence; Design methodology; Histograms; Iterative algorithms; Iterative methods; Nearest neighbor searches; Vector quantization; Voting;
Conference_Titel :
Information Theory, 1994. Proceedings., 1994 IEEE International Symposium on
Conference_Location :
Trondheim
Print_ISBN :
0-7803-2015-8
DOI :
10.1109/ISIT.1994.394628
