Research And Implementation Of Intelligent Alarm System For Telecommunication Networks

Modern telecommunication networks are characterized with large scale, it means that network management system can debug in time to make it sure that the network patency and service normal running. As an important problem in network fault management, alarm correlation analysis can help network administrators to delete redundant alarms, locate faults and predict faults before they happen. Data mining provide a effective method to exactly location the fault. Data mining is from large databases or data warehouses to extract knowledge of interest. That is from a large number of incomplete, random data to find out the useful information, its purpose is to obtain data from the massive effective, novel, potentially useful, end-users can understand the process of model. Data mining is a decision support process. It is mainly based on artificial intelligence, machine learning, statistical techniques, highly automated analysis of data from the original company to make inductive reasoning, from which excavated a potential model to help business decision-makers to adjust marketing strategies, risk reduction, to make relatively correct decision. However, traditional alarm correlation analysis methods can hardly work well when networks are complex and changeful, while the knowledge discovery method can overcome the shortage of traditional.This paper improve the Apriori algorithm for the Alarm data characteristics of telecommunications networks, proposed NCApriori algorithm,made it suitable for network alarm association rules mining, then import it into network alarm system, design a system for mining model framework, so that the entire network alarm system, through the association rule mining to provide a more meaningful fault diagnostic information. The Apriori Algorithm mainly from the following aspects, first, provide a priority pruning strategy does not make sense to avoid generating the frequent item sets; Secondly, k-1 dimensional item sets generated by connecting operation k item sets repeatedly scan the database to avoid the pressure brought; Finally, the rules generate alarms when applied to the data by adding constraints in order to improve the quality of generated rules. This is a fully applicable to the network alarm data, association rules improved algorithm not only has a certain degree of science forward-looking and challenging, but also have strong practical significance and maneuverability.This paper describes an improved algorithm for mixed weighted association rules, its alarm data for the characteristics of how those improvements carried out in detail. Meanwhile, the overall framework of the telecommunications intelligent alarm system design, as well as the association rule mining algorithm module a summary of Describes how to achieve according to the design of telecommunications intelligent alarm system, with the results show the potential of the system for mining the effectiveness of the rules, and conducted a quantitative analysis of experimental results.