Music can motivate many daily activities as it can regulate mood, increase productivity and sports performance, and raise spirits. However, we know little about how to recommend songs that are motivational for people given their contexts and activities. As a first step towards dealing with this issue, we adopt a theory-driven approach and operationalize the Brunel Music Rating Inventory (BMRI) to identify motivational qualities of music from the audio signal. When we look at frequently listened songs for 14 common daily activities through the lens of motivational music qualities, we find that they are clustered into three high-level latent activity groups: calm, vibrant, and intense. We show that our BMRI features can accurately classify songs in the three classes, thus enabling tools to select and recommend activity-specific songs from existing music libraries without any input required from user. We present the results of a preliminary user evaluation of our song recommender (called PepMusic) and discuss the implications for recommending songs for daily activities.

中文翻译：

---

PepMusic：日常活动歌曲的动机品质

音乐可以激发许多日常活动，因为它可以调节情绪，提高生产力和运动表现并振奋精神。但是，对于如何推荐具有特定背景和活动意义的歌曲，我们知之甚少。作为解决此问题的第一步，我们采用理论驱动的方法并对布鲁内尔音乐分级量表（BMRI）进行操作，以从音频信号中识别出音乐的动机品质。当我们通过激励性的音乐特质来观察经常收听的14项日常活动的歌曲时，我们发现它们被分为三个高水平的潜在活动组：平静，充满活力和强烈。我们证明了我们的BMRI功能可以准确地将歌曲分类为三个类别，因此，无需用户进行任何输入，工具就可以从现有音乐库中选择和推荐与活动相关的歌曲。我们介绍了歌曲推荐器（称为PepMusic）的初步用户评估结果，并讨论了推荐歌曲用于日常活动的含义。