The article explores neural networks. Artificial neural networks can also be successfully applied to Data Mining tasks. Artificial neural networks are simplified models of biological neural networks of the brain of living beings, and in these models with a large number of parallel working rather simple calculators "synapse (multiplier) adder threshold element" solves rather complex problems with high speed. Such tasks include classification, clustering, searching for patterns, associations, etc. In particular, self-organizing maps (Kohonen networks) improve the understanding of the data structure, which will help to more effectively conduct exploratory data analysis, discover new phenomena, etc.