الفهرس 
IntroductionOnly 14 pages are availabe for public view
 |  | from | 126 |  |  |
| | | from | 126 | | |
Abstract
Nowadays, Artificial intelligence is of interest to many scholars. It is a wide-ranging branch of computer science concerned with building smart machines capable of performing tasks that typically require human intelligence. It is a field that merges computer science and the robust data to solve problems. In addition to AI, there are sub-fields of machine learning and deep learning that seek to create expert systems which make predictions or classifications based on input data. This thesis introduces a problem of using support vector machine (SVM) to classify the dataset where, SVM model is used to separate between the classes by choosing hyperplanes that maximize the gab between these classes but how to control the misclassification error. The thesis introduces a proposed bi-objective support vector machine model to classify the datasets where, the first objective is to maximize the gab between classes and the second objective is to minimize the misclassification error then solve this optimization problem by using the weighting method. Then, it introduces a proposed fuzzy bi-objective support vector machine model by adding a fuzzy parameter to the previous model. Some concepts that help to identify the interactive approach such as multi-objective optimization, decision-making, interactive methods and interactive machine learning are introduced in this thesis. This thesis introduce two applications that use the proposed Bi-Objective Support Vector Machine where, the first application is related to corona virus disease and the second application is related to the pneumonia disease. The experimental evaluation was carried out using the dataset of Chest X-ray publicly available on the Kaggle. Finally, The computational results demonstrate the effect of the weighting parameters with the bi objective support vector machine on the efficiency of the Residual neural network model. The comparison study has been done between different types of Residual neural networks (18-50-101) plus BO-SVM proposed model at different training to testing ratio.