Graph Privacy Computation And Quantization Method Based On Structural Entropy

At present,The development of big data and 5G industry has enabled the rapid growth of various types and quantity of data in different fields.While the open sharing of data has gradually changed the way people live,learn and work,the problem of personal privacy security protection has become increasingly significant.Structure entropy is the real structural information of network structure decoded in the embedded large-scale noise structure,which can effectively measure the private information in the dynamic complex network.To a great extent,it makes up for the defect that information entropy can not measure the private information in the graph structure.Therefore,the research of structure entropy on graph structure is very important for realizing data protection and data value mining in large networks such as social networks.In this paper,information theory,structural information theory and data mining are used as the tools to study the privacy protection of structure entropy in graph structure.The content of the research involves graph privacy measurement method based on structural information theory,mutual information between graph structures and its application,achieving graph clustering privacy preservation based on structure entropy in social internet of things.The specific work is as follows:(1)Graph privacy measurement method based on structural information theory.Combining the definition of structure entropy in structural information theory,we first define a model graph G,and then use a segmentation algorithm to divide graph G into all possible modules as much as possible.Furthermore,we use the calculation formula for calculating the structural entropy in the structural information theory to calculate the K-dimensional structural information of the graph G,thereby preliminarily measuring the amount of information hidden by the private data in the graph structure and the leakage degree of private data.(2)Structural mutual information and its application.Firstly,introduce the structure entropy in concept of structural information theory,combining the definition of Shannon mutual information in information theory to build the structure of two or more graph.And the mutual information between calculation chart and graph,reflect the strength of the correlation degree between the two graph.Furthermore,we quantitative measure of the uncertainty between the graph and the privacy of data leakage degree.Finally,apply structural mutual information on social networks,effective measure of the network group and the correlation degree between degree of privacy.(3)Achieving graph clustering privacy preservation based on structure entropy in social internet of things.Firstly,homomorphic encryption is used to encrypt users' private data information in social networks,and the encrypted user data is regarded as graph nodes.Secondly,put forward the solving algorithm of two-dimensional structural information and clustering algorithm based on entropy reduction principle,the nodes in the graph structure get corresponding module and divided into K-dimensional structural information algorithm for further detailed clustering has been divided into modules.Then,after a complete graph clustering fully reflect the correlation degree between the nodes in figure graph.Finally,the normalized structural information and the similarity degree of network nodes are introduced to analyze the correctness and similarity degree of clustering results.