Hits:

Indexed by:期刊论文

Date of Publication:2022-06-28

Journal:系统仿真学报

Affiliation of Author(s):电子信息与电气工程学部

Volume:21

Issue:8

Page Number:2152-2155,2194

ISSN No.:1004-731X

Abstract:Improving the linear ridge regression by using the kernel function which meets the Mercer's condition, a novel nonlinear kernel ridge regression based on directly optimization was proposed. Compared with the support vector machine regression, this method is characterized by computing the squared lose function and the number of the parameters in the kernel ridge regression that need for selection are reduced, moreover, the time of modeling can reduce using the directly optimization. The simulation results and the application results in the soft sensing of aviation kerosene show the effectiveness of this method.

Note:新增回溯数据