Learning vector quantization models (LVQ) belong to the most successful machine learning classifiers. LVQs are intuitively designed and generally allow an easy interpretation according to the class dependent prototype principle. Originally, LVQs try to optimize the classification accuracy during adaptation, which can be misleading in case of imbalanced data. Further, it might be required by the application that other statistical classification evaluation measures should be considered, e.g. sensitivity and specificity like frequently demanded in bio-medical applications. In this article we present recent approaches, how to modify LVQ to integrate those sophisticated evaluation measures as objectives to be optimized. Particularly, we show that all differentiable functions built fro contingency tables can be incorporated into a LVQ-scheme as well as receiver operating characteristic curve optimization.
CITATION STYLE
Villmann, T. (2016). Sophisticated LVQ classification models - beyond accuracy optimization. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 10087 LNCS, pp. 116–130). Springer Verlag. https://doi.org/10.1007/978-3-319-50862-7_9
Mendeley helps you to discover research relevant for your work.