Invention Grant
- Patent Title: Method for the machine learning of frequent chronicles in an alarm log for the monitoring of dynamic systems
- Patent Title (中): 用于监视动态系统的报警日志中机器学习频繁编年史的方法
-
Application No.: US10554303Application Date: 2004-05-25
-
Publication No.: US07388482B2Publication Date: 2008-06-17
- Inventor: Christophe Dousson , Fabrice Clerot , Francoise Fessant
- Applicant: Christophe Dousson , Fabrice Clerot , Francoise Fessant
- Applicant Address: FR Paris
- Assignee: France Telecom
- Current Assignee: France Telecom
- Current Assignee Address: FR Paris
- Agency: Oblon, Spivak, McClelland, Maier & Neustadt, P.C.
- Priority: FR0350181 20030527
- International Application: PCT/FR2004/050205 WO 20040525
- International Announcement: WO2004/107652 WO 20041209
- Main IPC: G08B29/00
- IPC: G08B29/00
Abstract:
An automatic method for learning frequent chronicles in an alarm log of a dynamic system for supervision of the system, and a learning system for using the method in a supervision system. According to the method, in an alarm log, alarm sequences are selected and then grouped in groups of similar sequences. Alarms of the groups are then used to produce partial logs. Frequent chronicles are then learned on each transmitted partial log and the partial set of frequent chronicles is determined. Then a set of chronicles in the log is formed with chronicles of the various partial sets.
Public/Granted literature
Information query
