International Conference on Computer Technology and Development, 3rd (ICCTD 2011)
232 Weighted Manifold Multi-Plane Twin Least Squares Classification
Download citation file:
In this paper, we propose weighted Manifold Multi-plane Twin Least Squares classifier which concerns the local geometry of the samples by a manifold regularization term and measures the importance degree of each sample using K neighboring relation. In addition to keeping the respective advantages of TWSVM, LSTSVM and some graph learning algorithms, our methods improve the separation of the points sharing different classes. Also experimental evidence suggests that our methods are effective in performing classification task.