Biological network interaction construction method

The invention belongs to the technical field of deep learning big data medical treatment, and particularly relates to a biological network interaction construction method, which comprises the steps of establishing an undirected connection network graph G; carrying out primary processing on the undirected connection network graph G: aggregating information of neighbor nodes of different levels to a source node, and generating a wide-area source node representation vector; constructing a specific sub-graph for each source node from the original network graph G, and learning specific representation vectors of the source nodes; and based on the sum, predicting an interaction relationship between any two nodes. According to part of information, the existing interaction relation between the nodes can be accurately predicted, meanwhile, the unknown interaction relation can be predicted, and a direction is provided for follow-up biomedicine exploration. In addition, the interactive relationship deduced by the depth m