Learning non-parametric kernel via matrix decomposition for logistic regression

Kaijie Wang, Fan He, Mingzhen He, Xiaolin Huang

Abstract: Highlights•A matrix decomposition method is designed to pursue low-rank structure of the adjust matrix.•This method realizes the non-parametric adjustment of the kernel in logistic regression.•The idea of matrix decomposition is generalized to learn indefinite kernels, which further enhances the model flexibility.