Nonparametric Bayesian label prediction on a graph
This work addresses graph-based classification problems, likely for applications in network analysis or structured data, but appears incremental as it builds on existing Bayesian and graph methods.
The paper tackles binary classification on graphs by implementing a nonparametric Bayesian approach with a randomly scaled Gaussian prior that incorporates graph Laplacian geometry, proposing theoretically optimal and flexible variants, and demonstrates results on simulated and real data examples.
An implementation of a nonparametric Bayesian approach to solving binary classification problems on graphs is described. A hierarchical Bayesian approach with a randomly scaled Gaussian prior is considered. The prior uses the graph Laplacian to take into account the underlying geometry of the graph. A method based on a theoretically optimal prior and a more flexible variant using partial conjugacy are proposed. Two simulated data examples and two examples using real data are used in order to illustrate the proposed methods.