Skip to content

First Master Thesis on the topic of Graph Kernels and SVMs for Pattern recognition, focusing on low-complexity random walk kernels for graphs

License

Notifications You must be signed in to change notification settings

leoandeol/graph-kernels

Repository files navigation

Graph Kernels and Support Vector Machines for Pattern Recognition

This project was conducted during the first year of my master's at Sorbonne Université. The main focus was to find low-complexity methods for the random walk graph kernel, especially on labeled graphs.

Prerequisites

  • numpy
  • scipy
  • scikit-learn
  • grakel (to import graph databases)
  • control (dlyap)
  • slycot (necessary for dlyap in control)
  • jupyter (optional, for tests)

Content

  • Thesis report and presentation slides
  • A synthetic graph database generator
  • The Random Walk kernel introduced by (Vishwanathan et al, 2010)
  • 5 Acceleration Methods introduced in the same paper
  • Experiments on their speed and accuracy

About

First Master Thesis on the topic of Graph Kernels and SVMs for Pattern recognition, focusing on low-complexity random walk kernels for graphs

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published