Alarm Correlation Research and Implementation Based on Similar Data Sources
Yi Tang, Dahai Jin · 2016
In the modern society, static defect detection technology become an effective way to build reliable software.But the phenomena of misinformation and failing to report in the result of static defect detection makes testers needed to confirm the false alarms one by one manually.It is necessary to improve efficiency of alarm confirmation.In this paper, an alarm correlation method for similar data sources is proposed.The method uses data sources of the alarm to generate symbol informationthe define point information.With the symbolic information, we can generate a feature function which is consist of all alarm variable's DEF point symbols.According to feature function, we can generate an equivalent signature code.Finally, by comparing the similarity of the signatures to determine whether there is an association between the two alarm.Testers only need to pick up one or a few alarms in a class for confirmation after apply this technology which greatly improve the system's efficiency.