Abstract
In this paper, we developed two new classifiers: the kernelized geometrical bisection method and its extended version. The derivation of our methods is based on the so-called 'kernel trick' in which samples in the input space are mapped onto almost linearly separable data in a high-dimensional feature space associated with a kernel function. A linear hyperplane can be constructed through bisecting the line connecting the nearest points between two convex hulls created by mapped samples in the feature space. Computational experiments show that the proposed algorithms are more competitive and effective than the well-known conventional methods.
Original language | English |
---|---|
Pages (from-to) | 680-688 |
Number of pages | 9 |
Journal | Lecture Notes in Computer Science |
Volume | 4492 |
Issue number | 2007 |
DOIs | |
Publication status | Published - 2007 |