Skip to content

Latest commit

 

History

History
44 lines (29 loc) · 897 Bytes

README.md

File metadata and controls

44 lines (29 loc) · 897 Bytes
title layout
Heat kernel diffusion for local clustering
project

hkgrow: heat-kernel based local clustering

Kyle Kloster

David F. Gleich

These are research codes and may not work for you.

Download

Synopsis

compile % compile the mex files
G = load_graph('dolphins');
hkgrow_mex(G,seeds,t,eps);

Reusable codes

  • hkgrow_mex.m C++ MEX code for computing a set of best conductance via seeded heat kernel with input graph "A" and input parameters "seeds, eps, t".

Codes from others

  • pprgrow.m from personal page rank clustering

Results from the paper [TO DO: replace with KDD results]

To reproduce figure 2 (left), run:

test_tol_accuracy % generate the data
plot_tol_accuracy % plot the data