Application of Semi-Markov Processes for Prediction of Success of Musical Works

People tend to predict the success of their work, activities, inventions and so forth. This paper represents a methodology for predicting the success of musical pieces based on semi-markov processes. The method includes an algorithm for generating new compositions due to statistics from a chosen piece. Moreover, the study investigates an unproved statement about the convergence of accumulated random revenue to its expected value. After a brief review of the literature of the subject matter, mathematical and musical theory, the approach is described in two parts. The description is divided into two parts. The first one is about the algorithm for composing and its results. The second is about predicting success and investigating convergence. The technique and hypothesis are tested on the Waltz by A. Khachaturian from the music to Lermontov's drama ``Masquerade``.