الفهرس 
IntroductionOnly 14 pages are availabe for public view
 |  | from | 109 |  |  |
| | | from | 109 | | |
Abstract
This thesis concerned with a new way to diagnose the disease, which represents one of the major problems challenges facing Expert Systems, which are represented in the loss of data, which occurs during the process of diagnosis hypothyroidism disease by using Machine learning to complete loss of the attributes to make the classification process, also used Genetic Algorithms in the process of improving the efficiency of the system of Rule-based knowledge representation and construction of building an expert system to diagnose the disease called Diagnosis Expert System Shell. Two intelligent hybrid methods for diagnosing hypothyroid disease are introduced in this thesis in an attempt to address the above-mentioned issues. In the first approach, Machine Learning (ML) techniques are combined with expert systems in order to deal with the problem of environment changes. The well-known evolutionary classification ML algorithm, which is called C4.5, is employed in this method. A new Correlation-based Feature Selection (CFS) approach is introduced to the employed ML algorithm in order to cope with the KA problem. The predictive accuracy of C4.5 classifier can be improved due to the introduction of this proposed filter. The second hybrid model presents a new approach of building an expert system shell for disease diagnose corresponding to an evaluation for optimal features that allow system to diagnosis disease efficiently. In this model, genetic algorithm-machine learning method is introduced for discovering and improving the efficiency of the automated rule-based expert systems.