In recent years, there has been an increase in the number of artists who make use of automated music performances in their music and live concerts. Automated music performance is a form of music production using programmed musical notes. Some artists who introduce automated music performance operate parameters of the sound in their performance for production of their music. In this paper, we focus on the music production aspects and describe a method that realizes operation of the sound parameters via computer. Further, in this study, the probability distribution of the action (i.e., variation of parameters) is obtained within the music, using Bayesian filters. The probability distribution of each piece of music is transformed by passing through a Markov model. After the probability distribution is obtained, sound parameters can be automatically controlled. We have developed a system to reproduce the musical expressions of humans and confirmed the possibilities of our method.