| publication name | Extracting rules from trained neural network using genetic algorithm |
|---|---|
| Authors | Raafat A. El-Kammar, Atta E. El-Alfy, Mohamed I. Sharawy, Mohye- E. El-Alame |
| year | 2002 |
| keywords | |
| journal | Cairo University, Institute of Statistical Studies and Research, the Egyptian computer journal, 22-7-2002, Cairo, Egypt |
| volume | Not Available |
| issue | Not Available |
| pages | Not Available |
| publisher | Not Available |
| Local/International | International |
| Paper Link | Not Available |
| Full paper | download |
| Supplementary materials | Not Available |
Abstract
This paper presents a new algorithm for extracting accurate and comprehensible rules from trained neural network (ANN) using the genetic algorithm (GA). The new algorithm does not depend on the ANN training algorithms and it does not modify the training results. The GA is used to find the optimal values of the input attributes (chromosome),Xµ, which maximize the output function ψк of output node k. The function ψк= ƒ( χі, (WG1)ij, (WG2)j,ĸ) is nonlinear exponential function. The values (WG1)ij, (WG2)j,ĸ are the groups of weights between each input and hidden nodes, and each hidden and output nodes respectively. The optimal chromosome is decoded and used to get a rule belongs to class ĸ