Abstract
Music plays an important role in human history. Since music could influence the emotion and performance of people, it has been widely applied to enhance the efficiency of some particular tasks, such as sports or medical applications. For example, a jogger can run longer distance with a proper music. However, people usually select the music according to some personal preference. In this manner, it might not be able to reach the anticipative goal of task. Moreover, manually music selection is a time-consuming activity. Therefore, to select the proper music from a huge amount of music database became a major ibue. The aim of this study is to establish an intelligent music selection system. A method based on Music Information Retrieval (MIR) technique which can query and retrieve certain types of music was developed. This system was realized by application software. This system is not confined to appropriate music selection by tracing the user's heart rate variability. More importantly, it could keep users in a positive emotional state. This study expects to provide effective music patterns, contributing to the therapeutic music creation and the applications in the relevant field of music therapy.