A Novel Kernel Matrix Isomap Algorithm Based on Partial Least Squares
Bing Li, Feng Ming Guo, Yi He · Advanced materials research · 2012
In this paper, to alleviate the influence of the noise the inaccurate measurement in the complicated environment, based on the robust ability in multivariate linear regression of PLS, and in combination with nonlinear data dimension reduction of manifold learning, a novel kernel matrix Isomap algorithm is proposed. The contribution rate is used to find and delete the “short circuit” edge. The matrix constructed by double centered transformation and kernel transformation trick is mapped to a high dimensional feature space, finally the relative position is obtained by PLS. Compared with traditional Isomap and MDS, simulation results indicate that the algorithm has good topology stability, generalization property, robustness and lower computational complexity