The selection of alarm transaction set based on time window directly affects the effectiveness of alarm association rule mining

and the internal correlation of Automatically Switched Optical Network（ASON）alarm makes it have a certain clustering relationship.Therefore

we propose a alarm clustering and correlation method based on business time window selection.The multi-network alarm is first clustered according to the network business configuration.Then the time window is divided to generate a more strongly correlated alarm transaction set by determining the alarm growth period.Finally

we also investigate the alarm association rules.This method can adapt to the characteristics of multi-plane and multi-network alarm internal correlation of intelligent optical network

which has the advantages of effectively compressing the alarm transaction set

higher accuracy of alarm association rules

and support for subsequent fault diagnosis.The feasibility and effectiveness of this method are verified by the practical network data.