Abstract:Graphs are ubiquitous for describing and modeling complicated data structures, and graph embedding is an effective solution to learn a mapping from a graph to a low-dimensional vector space while preserving relevant graph characteristics. Most existing graph embedding approaches either embed the topological information and node features separately or learn one regularized embedding with both sources of information, however, they mostly overlook the interdependency between structural characteristics and node features when processing the graph data into the models. Moreover, existing methods only reconstruct the structural characteristics, which are unable to fully leverage the interaction between the topology and the features associated with its nodes during the encoding-decoding procedure. To address the problem, we propose a framework using autoencoder for graph embedding (GED) and its variational version (VEGD). The contribution of our work is two-fold: 1) the proposed frameworks exploit a feature-dependent graph matrix (FGM) to naturally merge the structural characteristics and node features according to their interdependency; and 2) the Graph Convolutional Network (GCN) decoder of the proposed framework reconstructs both structural characteristics and node features, which naturally possesses the interaction between these two sources of information while learning the embedding. We conducted the experiments on three real-world graph datasets such as Cora, Citeseer and PubMed to evaluate our framework and algorithms, and the results outperform baseline methods on both link prediction and graph clustering tasks.

Principal Component Analysis Based on Graph Embedding

Graph Learning in Low Dimensional Space for Graph Convolutional Networks

Graph Convolution RPCA With Adaptive Graph

Graph-Laplacian PCA: Closed-Form Solution and Robustness

Principal Component Analysis on Graph-Hessian

Sparse Projections over Graph.

Principal Component Analysis Based on Graph Laplacian and Double Sparse Constraints for Feature Selection and Sample Clustering on Multi-View Data

A graph regularized dimension reduction method for out-of-sample data.

Feature extraction based on graph discriminant embedding and its applications to face recognition

Graph Embedding with Constraints

Connecting Graph Convolutional Networks and Graph-Regularized PCA

GraphPCA: a fast and interpretable dimension reduction algorithm for spatial transcriptomics data

Multi-Dimension Principal Component Analysis based on face recognition

Projective Unsupervised Flexible Embedding with Optimal Graph.

Maximally Correlated Principal Component Analysis Based on Deep Parameterization Learning.

Graph optimization for unsupervised dimensionality reduction with probabilistic neighbors

Graph classification based on skeleton and component features

Feature-Dependent Graph Convolutional Autoencoders with Adversarial Training Methods

Perturbation analysis on PCA plus graph embedding methods and PCA plus exponential graph embedding methods

A THEORETICAL FRAMEWORK FOR MATRIX-BASED FEATURE EXTRACTION ALGORITHMS WITH ITS APPLICATION TO IMAGE RECOGNITION

Principal Component Analysis in the Graph Frequency Domain