Combining High Level Symptom Descriptions and Low Level State Information for Configuration Fault Diagnosis
Ni Lao, Ji-Rong Wen, Wei‐Ying Ma, Yi‐Min Wang · 2004
Automatic fault diagnosis is an important problem for system management. In this paper, we combine high level symptom descriptions and low level state information to solve the system fault diagnosis problem. We extract state-symptom correlation information from knowledge sources in text format, and then use symptom similarity to rank the candidate system states. We apply the method to Windows Registry problems to help Product Support Service (PSS) engineers. Promising results with two different knowledge sources show the robustness of our method. Finally, we explain why this combination is successful and also discuss its limitations.