Research And Application Of Fault Alarm Impact Analysis Based On Association Rules In Communication Network

As an integral part of network maintenance, network management increasingly causespeople’s attention. Because of the complexity and hugeness of the communication networks,network management becomes a problem of personnel network maintenance, especially in theface of vast amounts of alarm data. There will be some associations between equipments andlinks. There will also be a relationship that exists in each alarm which can be used as a sourceof alarm and leads to a variety of derivative alarms, and causes the network managementmore difficult.Based on the randomness and uncertainty of alarm management, this paper uses animproved incremental association rule mining algorithm to analyze the data, on the alert, andovercomes time-consuming on multiple scan of the database in the incremental associationrules algorithm. The impact of the fault alarm analysis is introduced, including correlationanalysis and the business and customer impact analysis. The theory and method of the twoparts are researched in-depth. The main impact of alarm analysis techniques and methods isdescribed in detail, using the method of knowledge discovery in alarm fault impact analysisby comparing the obtained methods.In this paper, on the basis research of network element alarms between the horizontalrules (correlation analysis), the three levels of longitudinal studies between the relatedinfrastructures, the users associated with the business relationship are increased. The alarmanalysis model based on incremental association rule mining impact analysis is designed.According to the impact analysis of fault alarm, correlation analysis model and business andcustomer impact analysis model are proposed. Combined with the advantages of two parts,model of the impact of fault alarm analysis is proposed, which used to influence the impactanalysis will greatly improve the analysis speed.Related experiments are carried out to analyze the impact of fault alarm on associationrule mining in this paper. Leading the raw data into the database, and preprocessing the data,and then incremental association rule mining, and alerts are correlated display, analysis offailure alarm correlation, association rule mining on the results for display and analysis.Finally, it is compared in association rules under different support, during the improvedalgorithm, the FUP algorithm and the Apriori algorithm, and also in execution time and thenumber of association rules in the new database, which proves that the improved algorithmmeets the requirements of practical application.