pyGPs is an object-oriented Python package for Gaussian processes for machine learning. Download pyGPs HERE or view it on GitHub: https://github.com/marionmari/pyGPs.
Coinciding Walk Kernel
The coinciding walk kernel is a kernel among the nodes of a graph. It can be used for node label classification. The implementation is available HERE.
Propagation Kernels are a framework to compute a kernel between graphs. They can be used for graph classification for labaled, attributed, unlabaled, and grid graphs. A MATLAB implementation is available here: https://github.com/marionmari/propagation_kernels. A Python implementation for labaled and unlabaled graphs is part of pyGPs or can be directly found HERE.