Knowledge discovery from communication network alarm databases

The technique of Knowlege Discovery in Databases (KDD) to learn valuable knowledge hidden in network alarm databases is introduced. To get such knowledge, we propose an efficient method based on sliding windows (named as Slidwin) to discover different episode rules from time squential alarm data. The experimental results show that given different thresholds parameters, large amount of different rules could be discovered quickly. Foundation item: Supported by the National 863 High-Tech Project (863-306-Z705-02) and National Natural Science Foundation of China (69896240) Biography: Wang Xi-miao (1973-), female, Ph. D. candidate. Research interests: network management and information processing.