Abstract
We present a novel approach for Fuzzy-Input Fuzzy-Output classification. One-Against-All Support Vector Machines are adapted to deal with the fuzzy memberships encoded in fuzzy labels, and to also give fuzzy classification answers. The mathematical background for the modifications is given. In a benchmark application, the recognition of emotions in human speech, the accuracy of our F2-SVM approach is clearly superior to that of fuzzy MLP and fuzzy K-NN architectures. © Springer-Verlag Berlin Heidelberg 2007.
Author supplied keywords
Cite
CITATION STYLE
Thiel, C., Scherer, S., & Schwenker, F. (2007). Fuzzy-input fuzzy-output one-against-all support vector machines. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 4694 LNAI, pp. 156–165). Springer Verlag. https://doi.org/10.1007/978-3-540-74829-8_20
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
