Fig. 9From: MSPCD: predicting circRNA-disease associations via integrating multi-source data and hierarchical neural networkOverview of our proposed MSPCD method for predicting circRNA-disease assoications. Firstly, it takes the similarity of circRNA i and disease j as input and outputs their high-order non-linear features through three fully connected layers. Secondly, we use the dot product to acquire the high-level interactive feature of i and j. The result and the high-order features are concatenated to generate a new vector fed into DNN to finally realize the association prediction between circRNA and diseaseBack to article page