A new nonparametric feature mapping technique for pattern classification is proposed and compared experimentally with a principal component and Sammon’s mapping methods. We use the mapped training-set vectors for an active weights initialization of the multilayer perceptron classifier in a two-variate mapped space. Simulations have shown a usefulness of the proposed weights initialization method for designing the perceptrons when we need to obtain highly nonlinear decision boundaries.
CITATION STYLE
Raudys, A. (1998). A nonparametric data mapping technique for active initialization of the multilayer perceptron. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 1451, pp. 989–996). Springer Verlag. https://doi.org/10.1007/bfb0033329
Mendeley helps you to discover research relevant for your work.