Development of parsimonious fuzzy neural network for data driven modelling
In many complex real-world engineering problems, it is well known that the optimal number of rules is unknown. Hence, the fuzzy systems need to acquire knowledge by continuous self-organizing its rules to adapt to changing patterns in evolving data streams. Much-efforts have been devoted by several...
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Theses and Dissertations |
| Language: | English |
| Published: |
2013
|
| Subjects: | |
| Online Access: | http://hdl.handle.net/10356/51025 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Nanyang Technological University |
| Language: | English |
| Summary: | In many complex real-world engineering problems, it is well known that the optimal number of rules is unknown. Hence, the fuzzy systems need to acquire knowledge by continuous self-organizing its rules to adapt to changing patterns in evolving data streams. Much-efforts have been devoted by several scientists to develop self-organizing fuzzy systems. Yet, the majority of conventional techniques are mainly concerned with independent issue on structural complexity, prediction accuracy, and structural complexity without taking into account a balance among them. Highly accurate methods usually suffer high computational burden thereby limiting scalability in practical applications. In this dissertation, two novel self-organizing fuzzy systems, which attempts to strike a balance among structural complexity, prediction accuracy, and training speed, are proposed. |
|---|