Research On Application Of Large Data Analysis Method In Optimizing Time Series Of Communication Network

Network optimization refers to the optimization of existing network resources,so that network quality tends to the ultimate requirement of network planning.It is a huge system engineering with large tasks,strict time requirements and complicated process.The characteristic of online operation network is that there are too many optimization points.How to formulate a reasonable network optimization schedule is a difficult problem in the current network optimization work.Big data analysis is a typical application-oriented technology.It can not only query and traverse past data,but also find out the potential relationship between past data,so as to promote the transmission of information.This paper mainly discusses the combination of network optimization engineering and large data analysis application mode,and studies how to apply large data analysis technology to network optimization system,so as to find out the network optimization sequence that meets the needs in large data.This paper first discusses the current situation and development of network optimization,and focuses on the methods and processes of network optimization.At present,the main method of network optimization analysis is based on the comprehensive analysis method of market development direction and network state analysis.Because the amount of analysis data is too large,network state ranking is the main method of network optimization analysis after comprehensive analysis,and the ability of time-sharing analysis for all users of the network is very weak.In this paper,by using the method of large data analysis and mining,we can deal with large sample data sets.By analyzing the algorithms and practical application forms of data mining,we combine them to study how to apply data mining to network optimization timing confirmation.In the whole process of network optimization data mining,the key work is the selection of basic data sets and the preparation of data quantification.The main contradiction to be solved is the abnormal disturbance caused by the large difference of data units in the key indicators of communication network.In order to improve the accuracy of the algorithm,the classical K-means convergence algorithm is adjusted and optimized by using the concept of ant colony.At the same time,in order to solve the problem of low efficiency of ant colony algorithm,this paper improves it based on genetic algorithm.The improved ant colony algorithm shortens the number of iterations and reduces the amount of computation.The clustering results are exactly the same as the standard clustering results,but the clustering effect is better than the basic ant colony algorithm.On the other hand,considering the data magnitude,this paper builds a distributed computing experiment platform with three computers through Distributed Computing Toolbox,which further improves the data mining model's ability to process large data sets.After continuous exploration and attempt,the eventual distributed ant colony convergence algorithm can more accurately solve the basic classification of a large number of network complaint index data,and provide reasonable suggestions and priority timing guidance for the huge network optimization work.Provide very favorable support for network optimization,user demand processing,and even market expansion direction of operating enterprises.