| 摘要 |
The present invention discloses an error log analysis system comprising a diagnostic unit and a training unit. The training unit includes a plurality of historical error logs generated during abnormal operation or failure from a plurality of machines, and the actual fixes (repair solutions) associated with the abnormal events or failures. A block finding unit identifies sections of each error log that are in common with sections of other historical error logs. The common sections are then labelled as blocks. Each block is then weighted with a numerical value that is indicative of its value in diagnosing a fault. In the diagnostic unit, new error logs associated with a device failure or abnormal operation are received and compared against the blocks of the historical error logs stored in the training unit. If the new error log is found to contain block(s) similar to the blocks contained in the logs in the training unit, then a similarity index is determined by a similarity index unit, and solution(s) is proposed to solve the new problem. After a solution is verified, the new case is stored in the training unit and used for comparison against future new cases. |
