Researchers analyzing large (> 100,000 objects) data sets with the methods of cluster analysis often face the problem of computational complexity of algorithms, that sometimes makes it impossible to analyze in an acceptable time. Common solution of this problem is to use less computationally complex algorithms (like k-means), which in turn can in many cases give much worse results than for example algorithms using eigenvalues decomposition . The results of analysis of the actual sets of this type are therefore usually a compromise between quality and computational capabilities of computers. This article is an attempt to present the current state of knowledge on the classification of large datasets, and identify ways to develop and op...