الفهرس 
AcknowledgementOnly 14 pages are availabe for public view
 |  | from | 32 |  |  |
| | | from | 32 | | |
Abstract
This thesis presents a Rule-based Interval State Machine System (RISMS) for monitoring, and analyzing the behavior of interval-based data. This system consists of two proposed layers. The first layer is the rule-based real-time data mining layer that contains three modules: data monitoring and behavior analysis module, the federated database module, and the acquisition module. Three algorithms as well as four Key Performance Indicator (KPI) attributes are introduced by the first layer for monitoring and analyzing the interval-based data. The second layer is the knowledge-based monitoring layer where actions and behaviors are controlled with well-defined and structured knowledge bases. Besides, this layer paves the road of using business intelligence paradigm for monitoring systems by introducing a generic rule-based web-oriented monitoring agent and server modules. Both modules are characterized by their customized rules for workflow visualization, monitoring and generating necessary alerts to the operators. Additionally, a unified data representation and communication model is adopted and prototyped together with an ontology-based knowledge management approach to support monitoring real-time data processing and supervising the various system resources. Implementation and experimental results have proven the validity and effectiveness of the monitoring system.