Skip to main content
WorldCist'18 - 6th World Conference on Information Systems and Technologies

Full Program »

System Evaluation of Construction Methods for Multi-class Problems using Binary Classirers

Construction methods for multi-valued classification (multi-class) systems using binary classifiers are discussed and evaluated by a trade-off model for system evaluation based on rate-distortion theory. Suppose the multi-class systems consisted of M(/geq 3) categories and N(/geq {M -1}) binary classifiers, then they can be represented by a matrix W, where the matrix W is given by a table of M code words with length N, called a code word table. Applying document classification data, the relationships between the probability of classification error Pe and the number of binary classifiers N for given M are investigated, and we show that the systems have desirable properties such as \Flexible", \Elastic", and so on. Additionally, hand-written alphabet recognition data are examined, and we show that the desirable properties are also supported.

Shigeichi Hirasawa
Waseda University
Japan

Gendo Kumoi
Waseda University
Japan

Manabu Kobayashi
Shonan Institute of Technology
Japan

Masayuki goto
Wseda University
Japan

Hiroshige Inazumi
Aoyama Gakuin University
Japan

 

Powered by OpenConf®
Copyright ©2002-2017 Zakon Group LLC