| Telecom operators as the basis for service providers of information services,quality and stability of the network is one of the important factors that affect the future of competition in the telecommunications network.How to improve operational efficiency and reduce costs become a problem on which operators focus.Meanwhile,telecom operators accumulated a large amount of historical data need large data mining tools to enhance data availability.This paper first introduces the large data processing related technologies,as well as the mainstream model of data mining techniques,methods and tools.Then,based on the data sharing platform operators building and network features,a set of data sharing platform operators indicator system was built.The system comprehensively considers the data management software capabilities,the physical platform and the run-time data in the data sharing platform.Then we propose a solution based on large data mining technology Hadoop and R language,designed a large data mining platform.Combining Hadoop distributed computing power handling capacity and the matrix calculating capacity of R language,this architecture achieves efficient mass data mining.On this basis,we study the K-means algorithm and Apriori and its realization based MapReduce parallel,respectively,experiments show that the algorithm has good efficiency and scalability.Finally,we propose a Gn xDR data analysis methods and establish a analytical models,from a business point of view of the quality of data clustering user data,and achieve frequent itemsets mining district level and user-level exception scenarios discovery,excavation data traffic surges,sags from the perspective of data quality through dynamic association rules,association data quality degradation scenarios,the method can effectively detect potential problems in the network factors for upper applications targeted mutation problem solving to provide reference data.This paper combines big data technology and data mining techniques to analyze the data for the current network.The results enable network managers to understand the operation of the network status and dynamic trend,guide and optimize network operation and maintenance,provide a strong support to enhance network operation and maintenance management and rich the operational analytic hierarchy. |
