Intelligent Analysis of Music's Affective Features and Expressive Pattern

Affective experience is the artistic charm that music brings to people. Existing studies have shown that people' preferences for affective features of music are significantly different and greatly affected by their psychological and physical status as well as the environment scenes where they are. Therefore, how to accurately extract and analyze music's affective features is of great significance for musical evaluation, appreciation, and composition, as well as for finding the proper music to match with people's preference best. This paper firstly built a group of acoustic parameters to extract the affective features of music, and then converted those parameters into the variables' values of Thayer model through machine learning. On this basis, the expressive pattern of music's affective features was described by the time series of the above affective value's changes. Experimental result showed that the proposed method provided an effective intelligent method for analyzing the affective features of music. Superiorly, it conducted the expressive pattern to reflect the dynamic changes of music's affective features.