Telecommunication fraud security federal detection method fusing homogeneous graph and bipartite graph

The invention relates to a secure federal telecommunication fraud detection method fusing a homogeneous graph and a bipartite graph, and belongs to the field of big data analysis and mining. S1, based on user service data of a telecommunication operator, extracting and preprocessing voice call data, short message communication data and mobile phone application access data of a user; s2, constructing a telecommunication user social network homogeneous graph and user mobile phone application bipartite graph data set; s3, constructing a homogeneous graph embedded network according to the homogeneous graph of the social network, constructing a bipartite graph embedded network according to the bipartite graph applied to the mobile phone accessed by the user, sampling user nodes to obtain a neighbor node co-occurrence sequence, carrying out iterative training to obtain the embedded representation of each node, and fusing the embedded representation as the embedded representation of the user; and S4, different participants extract local telecommunication user features according to local data characteristics, perform joint training on local data of different institutions by using a security federal gradient boosting tree classification model, and output a final prediction result of the fraud number.