Proposal of method for generating musical compositions of different genres

Abstract

The work is dedicated to the research and development of methods and algorithms that automate and support the process of technical creativity, from the point of automated creation of musical compositions with different genres. The method is based on the use of neural networks to predict composition, and also suggests using various models that have been trained by songs in certain genres to improve the quality of the resulting musical composition. This work describes the algorithm itself, the process of collecting and classifying data for training, the process of training a neural network. In addition, it describes the process of choosing a neural network architecture. Also, the work proposes a system architecture for the automated generation of compositions using the proposed algorithm.