الفهرس Only 14 pages are availabe for public view
 |  | from | 109 |  |  |
| | | from | 109 | | |
Abstract
Neural network is an important field of artificial intelligence research, due to its ability to learn complex functions which prove difficult with traditional systems. The success of neural networks in several application make it very hot topic. Although some of the results that can be achieved through the application of neural networks are remarkable, there has not been a rush to accept neural networks as a valid problem solving technique in the real world, this is because neural networks will output a result, but do not explain how they achicwd it. L’his is a t>iy piv>l>lew with the use of neural networks in medical diagnosis, and the main reason why they have not been widely accepted by the medical world. This can be eliminated by extracting a set of production rule that explain how the result was reached. There are several researches in this area try to arrive to more comprehensible rules with more accuracy.
C4.5 Rule-PANE is a machine learning algorithm, which could generate rules with strong generalization ability and strong comprehensibility. C4.5 Rule-PANE stands for ’C4.5 Rule Preceded by Artificial Neural Network Ensemble’. The algorithm uses an ensemble of neural networks as a pre-process to the C4.5 Rule induction process. Where, the generalization ability is provided by the neural network ensemble, whilst the comprehensibility is achieved through the application of the C4.5 machine learning algorithm.
The C4.5 Rule-PANE algorithm and its modification are applied over set of data, and compares its performance to a number of other machine learning techniques, including C4.5 Rule and Artificial Neural Network (ANN) and the results was discussed. Then the extracted rules from C4.5 Rule-PANE used to build the knowledge base of medical diagnosis expert system. The motivation for creating hybrid systems is to eliminate the drawback of each components, also