For the UNB-GCN  classification model, UNB as the input feature vector to the graph structure node, phenotypic information, such as gender, age and APOE genetic information, as the weight of the edges between nodes.