User Behavior Analysis Of Mobile Core Network Data

At present,the analysis of mobile communication core network data is mostly based on the application data such as voice service data from the carrier.However,research on data mining and analysis of IP data is less.With the rapid development of mobile communication networks and the researches on the IP-based process,the analysis of the IP data is necessary for mobile communications.At present,most of the methods for IP data mining and analysis are aimed at the Internet data.The process of the mobile communication core network data and the analysis of the user behavior need to be studied further.Aimed at the instant messaging data in the core network of mobile communication,this paper analyzes the characteristics of this type of data and establishes a filtering model based on the data characteristics to obtain the target data set.According to the importance,the relevance and the similarity of the data set,the behavior analysis of the users are found.The main contributions and innovations of this paper are as follows:1.Building an instant messaging data filtering model.This paper analyzed the data features based on the XMPP protocol and establied a data filtering model,and then the data filtering model is applied to the Hadoop platform to provide a unified data structure.This structure is used for the network establishment and data analysis.2.Critical node filtering based on node importance analysis.Aimed at the messaging data based on the XMPP protocol type,the visualization results are generated according to the connection relationship among the nodes,and the connection structure between nodes in the visualization result is classified.The PageRank algorithm is used to calculate the importance of each node.Combined with the summarized connection structure,the importance of nodes is further analyzed to discover key user nodes.3.User potential behavior analysis based on the inter-node relevance and similarity.Based on the analysis of the importance of nodes,this paper proposes a simplified bipartite graph model for the network structure of key users,and calculates the nodes using the correlation and similarity algorithms.At the same time,it also correlates the corresponding and the similarity of the corresponding nodes in the simplified model.Combining the two results of the calculation for the relevance and similarity of node trends.