Research On Estimation Algorithm Of Distribution Of Network Data And Analysis

With the development of era,the Internet has begun to infiltrate every aspect of people daily lives.From information inquiry to medical care and education,the Internet has provided the enormous convenience for the people.The development of the Internet has also led to explosive growth of network traffic data.Network traffic data records all the information of the surfer process.The analysis of network traffic data has gradually become hot point of research.Network traffic data analysis not only provides a basis for user behavior analysis,but also has important significance for network management,network structure optimization and abnormal traffic analysis.This thesis starts with the distribution of network traffic data.Firstly,we review the research status of distribution of network traffic data.Then,we introduce the Hadoop distributed platform and Spark computing framework,which provide the basis for the storage and calculation of mass network traffic data.This thesis introduces the existing skewed data distribution model and points out the shortcomings of these models.Then,a new distribution model,the Discrete Gaussian Exponential(DGX)distribution,is introduced,and the modeling process and evaluation method are introduced in detail.Next,the thesis models the data of network traffic with DGX distribution.It is found that there are a lot of data in the network which obey the DGX distribution.This thesis analyzes the traffic data on the basis of the DGX model.Finally,this thesis establishes a website diagram based on shared user,and models the data in the diagram with DGX model.This thesis studies the distribution of data in real Internet traffic,and analyzes these data on the basis of them.This is of great significance for user behavior analysis and network structure optimization.