On extracting knowledge from neural networks
The primary contribution of the thesis is an algorithm that overcomes the significant limitations of previous methods by taking an approach to the task of extracting comprehensible rules from trained networks. This algorithm is called HyperNeuro. Given a neural network, HyperNeuro extracts linear if...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis |
| Published: |
2008
|
| Subjects: | |
| Online Access: | http://hdl.handle.net/10356/3372 |