| Article ID | Journal | Published Year | Pages | File Type |
|---|---|---|---|---|
| 10321971 | Expert Systems with Applications | 2013 | 8 Pages |
Abstract
⺠A new sparsification method based on the significance measure of the training sample is proposed. ⺠The algorithm is updated by recursive least square algorithm in an online fashion. ⺠Sparsification and system updating are unified in a same framework. ⺠Deterministic convergence in kernel weight error vector is guaranteed. ⺠The computational cost is low and it is suitable for real-time applications.
Related Topics
Physical Sciences and Engineering
Computer Science
Artificial Intelligence
Authors
Haijin Fan, Qing Song,