Moment based estimation of stochastic Kronecker graph parameters
David F. Gleich, Purdue University
Art B. Owen, Stanford University
These codes are research prototypes and may not work for you. No promises. But do email if you run into problems.
Download
-
kgmoments.tar.gz (updated 2012-01-16) {: .nobullets}
Prereqs
- MatlabBGL
- mcode
- A working C/C++ compiler
- A working Matlab mex compiler
- Optional SNAP Codes (for liklihood calculations)
Instructions
Will be prepared upon request.
Overview
The package is organized by directory
matlab- All of the main matlab codes
data- precomputed data for the experiments
experiments- implementations of the experiments in the paper
initial- initial versions of many of the codes
skcoin- a python version of a coin-flipping random graph Kronecker graph generator
snap- snap code to compute the log-likelihood of a kronecker graph
web- this information and all the figures
Figures
|Experiment|Description|Figure|
|:------------------|:------------------------------------|:------------------|
|experiments/fitting/objective_fits.m | compute data for table on objective functions | |
|experiments/fitting/objectives_table.m | output table | Tab. 2 |
|experiments/fitting/kronecker_fits.m | compute data for table on kronecker parameters | |
|experiments/fitting/kronecker_fits_table.m | output table | Tab. 3 |
|experiments/fitting/kronecker_partial_fits.m | compute data for table on fitted kronecker parameters without certain features | |
|experiments/fitting/partial_fits_table.m | output table | Tab. 4 |
|experiments/identifiability/kron_identify_conflip.m | compute data on variance of kronecker fits to kronecker model | |
|experiments/identifiability/parameter_variance.m | output variance figures on a,b,c | Fig. 2 |
|experiments/identifiability/feature_variance.m | output variance figures | Fig. 3 |
|experiments/identifiability/feature_misfit.m | output misfit figures | Fig. 4 |