Abstract
This paper presents two networks, in which one solves least absolute deviations problems and another solves related problems and shows that they globally converge to exact solutions. The proposed neural networks have smaller size than existing neural networks [9], and do not have difficulty in selecting penalty parameters, in contrast to existing neural networks [11].
| Original language | English |
|---|---|
| Pages (from-to) | 13-21 |
| Journal | Neurocomputing |
| Volume | 19 |
| Issue number | 1-3 |
| DOIs | |
| Publication status | Published - 21 Apr 1998 |
| Externally published | Yes |
Research Keywords
- L1-norm minimization
- Neural networks
Fingerprint
Dive into the research topics of 'Neural networks for solving least absolute and related problems'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver