dssdyx / spectral Goto Github PK

Spectral Clustering
-------------------
Tim Nugent 2014

Performs K-means clustering using top K eigenvectors of an affinity (kernel) matrix. Kernel matrix is generated using the radial basis function kernel before being normalised. Input is an nXp data matrix in CSV format via the read_data function, or an nXp Eigen MatrixXd object via the constructor.


Build
-----

Install Eigen, e.g.:

sudo apt-get install libeigen3-dev

or download from http://eigen.tuxfamily.org/.

Make sure the include path is set right in the Makefile.

Compile with 'make' or:

g++ -O3 --std=c++11 -Wall -Wextra -I/usr/include/eigen3 src/*.cpp -o bin/sc

On OS X using brew:

brew install eigen

Then:

g++ -O3 --std=c++11 -Wall -Wextra -I/usr/local/include/eigen3 src/*.cpp -o bin/sc


Run
---

Run with 'make test', or:

bin/sc

Spirals test CSV file is read from the data directory.


Output
------

bin/sc
Read data/spirals.csv
Iterations 1 : Error 1.7881 : Error delta 1.7881 : Centroid movement 0.0711
Iterations 2 : Error 0.4435 : Error delta 1.3446 : Centroid movement 0.0197
Iterations 3 : Error 0.2425 : Error delta 0.2010 : Centroid movement 0.0122
Iterations 4 : Error 0.2130 : Error delta 0.0295 : Centroid movement 0.0000
Iterations 5 : Error 0.2130 : Error delta 0.0000 : Centroid movement 0.0000
Wrote data/spirals_clustered.csv
Cluster assignments:
1 1 2 2 1 2 2 2 1 2 2 1 1 2 2 1 1 1 1 1 2 2 1 2 2 2 2 1 1 1 2 1 2 2 1 2 1 2 1 1 2 2 2 2 1 1 1 1 1 2 1 2 1 1 2 2 2 1 1 1 1 1 2 2 1 2 1 1 1 2 2 1 2 2 2 1 1 1 1 2 1 2 1 2 1 1 1 1 1 1 2 1 1 2 2 2 1 2 2 2 2 1 1 1 2 2 1 2 2 2 1 2 1 1 1 1 2 2 1 1 2 1 1 1 2 1 2 1 1 1 1 1 2 2 2 2 2 1 1 1 2 2 1 2 1 1 1 2 2 2 1 2 2 2 2 2 2 1 1 1 1 2 1 2 1 1 1 2 1 1 1 1 2 1 2 1 1 1 2 2 1 1 1 2 2 2 1 1 2 2 2 2 2 2 2 2 1 1 1 1 2 1 2 2 1 2 1 2 2 2 2 2 1 2 1 2 1 2 1 1 1 2 2 2 2 1 1 1 2 1 1 2 2 2 2 2 1 1 2 2 2 2 1 1 1 2 2 2 2 2 2 1 1 2 2 1 1 1 1 1 1 1 2 2 2 2 1 2 1 2 1 1 2 2 2 1 2 1 1 1 1 2 2 1 2 1 2 2 2 1 2 1 2 2 1 1 2 2 2 1 
Plotting data using R...
src/plot.R

If R is installed, see spirals_clustered.png. A new CSV file (spirals_clustered.csv) is also written to data/ containing the cluster assignments.


To-do
-----

-Add more kernels
-Add automatic gamma tuning for RBF kernel


References
----------

K-means code adapted from:
https://github.com/pthimon/clustering

Useful info:
http://charlesmartin14.wordpress.com/2012/10/09/spectral-clustering/


Bugs
----

[email protected]