| Intrusion Alarm System is a subsystem of security system, contained detector, transmission line and controller mainframe. System can record alarm data when is running, and it is the Date Mining wanted:a large number of data. Through Data Mining, discovering system inner link of device type, monitoring environment and the rate of false alarm.System include a lots of detectors, what is having a low costing and property, so it lead to some projects'quality is bad too. Alarm system for a sub system of security system is not used. So the association rule can help us to find system information, it can guide designer, and boost the efficiency and safety, and help the system to protect us batter.Do some research in this project, we discovering system inner link of device type, monitoring environment and the rate of false alarm. We find the most important things of the rate of false alarm is environment. According to the different of environment, we can choose the most comfortable detector, it can obviously reduce the rate of false alarm.During the data preprocessing, we make addition of constraints for data, it can obviously boost the speed of the data preprocessing. In the process of discovering the association rule, we build a data cube and use index to find the frequent itemsets, then we use correlation analysis to filter the not interesting rules.Through the data mining, we find the rules what can not find by hummer, and also discover environment, time having effect in rate of false alarm. We find a new viewpoint to analyse the data of Intrusion Alarm System, discover the answer of reduce the rate of false alarm. |