Rectifying Self Organizing Map (RSOM)

For More Detail Visit: Project Page .

Implemented and designed by Eren Golge for the work "Gölge, E., & Duygulu, P.. ConceptMap:Mining noisy web data for concept learning , The European Conference on Computer Vision (ECCV) 2014."

RSOM is an algorithm as an extension of well-known Self Organizing Map (SOM). It mimics SOM clustering and additionally detects outliers in the given dataset in the cluster level or instance level.

Example call below with different commented call alternative.

There are two different implementations for training. Scipy implementation is designed for small scale problems and Theano version is for large scale problems which GPU utilization might help.

You can use this code for only SOM clustering as well. Up to my knowledge, it is the only SOM library designed for large data-sets in Python.

from sklearn import datasets
import time

data = datasets.load_digits().data

som = SOM(DATA = data, alpha_max=0.05, num_units=100, height = 10, width = 10)
#som.train_batch(100)
#start = time.time()
#som.train_stoch_theano(10)
som.train_batch_theano(num_epoch=100)
#som.train_stoch(10)
#clusters = som.ins_unit_assign
#print clusters
#stop = time.time()
#
print som.unit_saliency

#som_plot_scatter(som.W, som.X, som.activations)    
#som_plot_outlier_scatter(som.W, som.X, som.unit_saliency, som.inst_saliency, som.activations)
#som_mapping = som.som_map()
#som_plot_mapping(som_mapping)
print "Demo finished!"
#print "Pass time : ", stop - start