Thesis (Ph.D.) - Cairo University - Faculty of Computers and Information - Department of Computer Science

Multi-dimensional data sets characterized many fields which make the process of machine learning is more complex. Machine learning uses mathematical models, heuristic learning which provides controllability, observability, stability and easy updating process. In this thesis, the problem of reducing number of features and the number of generated rules are investigated. Rules dimensionality, the difficulties of interpretation and the rule behavior are issues that restrain efficient benefit of extracted knowledge. The main objective of this work is to develop a visual data mining model for automatically extracting and rendering reduced rules. The model aims to help decision maker for better understanding the discovered rules. The proposed model is designed to reach reduced number of features and visualized reduced number of rules with optimal performance

Issued also as CD