Research Article
Multiple Kernel Spectral Regression for Dimensionality Reduction
Algorithm 1
MKL-SR algorithm.
The training procedure of MKL-SR | Input: A set of training data, matrices and , a set of base kernels | and parameter in (21). | Output: Sample coefficient vectors ; | Kernel weight vectors ; | Make an initial guess for or ; | Computing the largest generalized eigenvectors of eigen-problem (2); | For do | (1) Compute ; | (2) A is obtained by solving the least squares problem (21); | (3) Compute in (24) and in (25); | (4) is optimized by solving optimization problem (24) via SDP; | Return A and ; |
|