Privacy Protection Of Weighted Network Graph Data Based On Differential Privacy

The emergence of various network applications has brought great convenience to people's life.At the same time,' the privacy of users has been more and more concerned.Network data is usually represented by graph,such as social network graph,business trade network graph,etc.These graphs may be composed of nodes,edges and the weights of edges,and they contain a lot of sensitive information.It is necessary to take appropriate privacy protection measures before the graph is published.At present,in the privacy protection of network graph data,either only the privacy protection of nodes,edges and other structures,or only the privacy protection of edge weight of the graph,almost all of them protect the privacy of graph data unilaterally.Therefore,this paper uses the privacy protection model of differential privacy to protect the privacy of the weighted network graph data,and proposes the corresponding effective algorithm to protect the privacy of the edge weight and structure of the network graph data at the same time.Firstly,this paper proposes a privacy protection algorithm(WGPA)with weighted graph.The protection of edge weight is prior to the protection of graph structure,and they are related and influenced each other.The Laplace mechanism of differential privacy is used to disturb the edge weight in the graph data set.The privacy budget is allocated to the edge weight sequence in each graph,and a 'reasonable privacy budget allocation strategy is designed.Considering the complexity of the structure of the graph,the graph is transformed into a coding pattern,and the disturbed weight value is taken into account in the coding process of the graph.After that,frequent subgraph mining algorithm is used to disturb and filter the structure of the graph.In the process of mining,Laplace mechanism and exponential mechanism of differential privacy are used to disturb and filter the structure of the graph,and finally the most ideal noise version of weighted network graph data is obtained.Therefore,this algorithm is suitable for privacy protection of weighted network graph data set,which can protect the privacy of users and the relationship between users.Then,a graph generation algorithm is proposed to disturb graph set.In the process of graph generation,Laplace noise is added to disturb the edge frequency,and the graph is formed by adding one edge at a time,and the edge set of candidate neighbors is formed.In order to select the neighbor edges that may be generated,this paper designs a reasonable neighbor edge filter condition,and finally outputs the disturbed graph data set.Besides,the whole privacy budget is reallocated,and the algorithm is applied to WGPA algorithm,and proposes the algorithm of graph generation and protection.Finally,the algorithm is verified by experiments.The real graph data set is used to verify the effectiveness of the algorithm.The algorithm proposed in this paper is compared with other algorithms in terms of data utility and other aspects.The experimental results show that the algorithm is effective.