Title :
A New Method for the Design of k-Class Bayes Classifiers
Author :
Doost, Roghayeh ; Rahmati, Mohammad ; Sayyadian, Abolghasem ; Shamsi, Hossein
Author_Institution :
Univ. of Amirkabir, Tehran
Abstract :
In this paper a new method for the design of K -class bayes classifier is proposed. As a benefit, the proposed method has less complexity rather than its conventional counterparts, OVA and OVO. Simulation results show that the classification error of the proposed method is approximately identical to the classification error of conventional approaches, OVO and OVA.
Keywords :
Bayes methods; error correction codes; pattern classification; Bayes classifiers; error correcting codes; k-class classifiers; Algorithm design and analysis; Design engineering; Design methodology; Error correction codes; Hamming distance; Support vector machine classification; Support vector machines; Voting; K -class classifier; OVA; OVO; bayes classifier;
Conference_Titel :
Electronics, Circuits and Systems, 2007. ICECS 2007. 14th IEEE International Conference on
Conference_Location :
Marrakech
Print_ISBN :
978-1-4244-1377-5
Electronic_ISBN :
978-1-4244-1378-2
DOI :
10.1109/ICECS.2007.4510946
