kerndwd: Distance Weighted Discrimination (DWD) and Kernel Methods
A novel implementation that solves the DWD and the kernel DWD which can be hundred times faster than the existing implementation of DWD based on the second-order-cone programming (SOCP). The package also handles some other general loss functions including large-margin unified machines (LUM).
||graphics, grDevices, stats, utils
||Boxiang Wang, Hui Zou
||Boxiang Wang <boxiang at umn.edu>