Research of outlier mining based on association rules applied in city operation system

How to find these symptoms in the mass monitoring data of the City Operational System is crucial and important to prevent the accidents happening. The association rule mining method is used to mine rules of infrequent itemsets and more interesting ones, and the outlier condition threshold is set upon the expert experience, searching for qualified data sets to distinguish the exceptional data among the monitoring ones of the City Operation System in this paper. The potential symptoms are acknowledged then through interactions of experts and machine, providing a useful decision-making support for the monitoring and preventing accidents happening. The algorithm proved in this paper has been applied in Beijing City Operation Administration Software System successfully and testified by practice well.