Authors

Baranov M. A.

Degree	Post-Graduate Student, Chair of Cybernetics, Moscow State Institute of Electronics and Mathematics, National Research University «Higher School of Economics»
E-mail	thenorthcat@gmail.com
Location	Moscow
Articles	Modification of the greedy clustering algorithm There are many clustering methods using different approaches to the cluster analysis problem. This paper deals with the greedy cluster algorithm modification. The essence of the modification is that the decision on whether to add a new document to the cluster or not is made on the basis of its similarity to the previously included in a cluster documents. The different optimization versions are discussed. The computing results are presented. Read more... Parallel greedy clustering algorithm The article is devoted to the development the parallel version of clustering algorithm based on the so-called greedy approach. Technology CUDA developed by NVIDIA is chosen for algorithm parallelization. The code and computing results for the similarity matrices of different sizes are presented. Read more...