The code for the pcv algorithm is in the directory
>	implementation/code/

The documentation of how to use the pcv algorithm practically is in a long
comment at the top of main.py.

To run the experiments first create a copy of the file SETUP_matlab_EXAMPLE.m
with the name SETUP_matlab.m where you enter your local directories.

Then open Matlab in the directory
>	implementation/experiments
and on the Matlab command line enter the command
>	setpath

Now you have access to the four algorithms that we used with the following
commands:
- pcv:			proj_wrapper
- message:		message_wrapper
- lmf:			lfm_wrapper
- lim:			bicluster_kmeans

The synthetic data is in the directory
>	implementation/experiments/data/
When you unzip one of the files, you will find the data that was used for the
experiments. Each directory will have size of approximately 500 MB. Inside the
lowest level folders, you will find the right clusters (...-patterns.csv), the
left clusters (...-leftClusters.csv) and the data (...-data.csv). The filename
indicates what parameters were used when the data was generated.  The
directories also contain a file "results.mat" which contains the outputs of the
algorithm when we ran the algorithms.

If you want to run the experiments, enter the following commands in Matlab:
- vary_p:		run_experiments('data/vary_p',8,8,0.425,10,10);
- vary_size:	run_experiments('data/vary_size',8,8,0.1564,10,10);
- vary_l:		run_experiments('data/vary_l',8,8,0.1906,10,10);
(In order to run vary_size_bad_param yourself, you'd have to change the
 parameters in run_experiments.m. We do not recommend this. You'd have to make
 changes in lines 112--130.)

If you just want to see the results of the experiments, use the following
commands:
- vary_p:				load_experiments('data/vary_p', '$p$')
- vary_size:			load_experiments('data/vary_size', 'Size of right clusters')
- vary_l:				load_experiments('data/vary_l','$\ell$')
- vary_size_bad_param:	load_experiments('data/vary_size_bad_param','Size of right clusters')

If you want to have the code for the heuristic please send me an email.
