PROGRESSIVE 2-CLASS DECISION CLASSIFIER FOR OPTIMIZATION OF CLASS DISCRIMINATIONS

A progressive two-class decision classifier (pTCDC) ir presented in th is article. This multilayer procedure converts a multiclass classifica tion problem into a several independent two-class separations. It not only provides the general advantages of hierarchical classification sc hemes over single-stage classification but it is also free of the need for hierarchical structure design and offers an optimal class pair di scrimination environment. At each decision node, only one class pair i s considered. Data processing aimed at maximising individual class pai r separations, such as feature selection, classification algorithm sel ection and data source selection or data transformation, be comes more reliable and efficient. Experiments carried out using an AVIRIS data set are presented and the results demonstrate that fewer features are needed and classification accuracy is improved with the new procedure compared with single-stage classification. (C) Elsevier Science Inc., 1998.