Application of machine learning in the development of databases of recommendation resources and music services

The problem of using computer-assisted learning technologies for organization musical recommendations according to user’s preferences. The main content of work is an analysis of spectrograms of musical compositions of various styles and directions and its further processing using the algorithms of neural network. In this research the problem of making new way of musical classification is solving and proving the idea about the fact that musical preferences can be developed on basics of compiling several packs of digital tags, that were stated to compositions because of the work of neural network.