Multiple meta paths combined for vertex embedding in heterogeneous networks

In the real-world many complex systems exist in the form of heterogeneous networks. As we all know, heterogeneous networks consist of various types of vertices and relations, so it is difficult to deal directly with data mining. At present, although many state-of-the-art methods of network representation learning have been developed, these methods can only deal with homogeneous networks or lose information when handling heterogeneous networks. In order to compensate for the weakness of the previous methods, we propose a multiple meta paths combined embedding (MMPCE) model to represent the heterogeneous networks. This method can automatically obtain the low-dimensional vector representation of vertices and preserve the rich semantic and structural information in the network. We conduct experiments on two real world datasets. The experimental results demonstrate the efficacy and efficiency of the proposed method in heterogeneous network mining tasks. Compare to the previous method, our model can cover a wider range of semantic information and be more flexible and scalable.