Recommending music to bar customers is an extremely difficult task without considering the mood of the bar environments. This paper proposes a multimodal mood detection mechanism that provides useful mood data for recommending music in bars. The proposed mechanism utilizes various sensor data representing alcohol concentration in the air, noise level, brightness level, crowd density, etc. to quantify the mood of the place environment using a simple mixed weight-based calculation.
The resulting representations can be used to categorize or rate the mood in a bar, which facilitates the development of music recommendation systems for bar environments.