Adapting self-organizing Maps to the MapReduce programming paradigm
Christian Weichel · Oldenbourg Wissenschaftsverlag eBooks · 2010
We present an adaption of the self organizing map (SOM) useful for cluster analysis of large quantities of data such as music classification or customer behavior analysis. The algorithm is based on the batch SOM formulation which has been successfully adopted to other parallel architectures and perfectly suits the map reduce programming paradigm, thus enabling the use of large cloud computing infrastructures such as Amazon EC2.