Authors: T Villmann M Kaden W Hermann M Biehl
Publish Date: 2016/08/27
Volume: 33, Issue: 3, Pages: 1173-1194
Abstract
This paper proposes a variant of the generalized learning vector quantizer GLVQ optimizing explicitly the area under the receiver operating characteristics ROC curve for binary classification problems instead of the classification accuracy which is frequently not appropriate for classifier evaluation This is particularly important in case of overlapping class distributions when the user has to decide about the tradeoff between high truepositive and good falsepositive performance The model keeps the idea of learning vector quantization based on prototypes by stochastic gradient descent learning For this purpose a GLVQbased cost function is presented which describes the area under the ROCcurve in terms of the sum of local discriminant functions This cost function reflects the underlying rank statistics in ROC analysis being involved into the design of the prototype based discriminant function The resulting learning scheme for the prototype vectors uses structured inputs ie ordered pairs of data vectors of both classes
Keywords: