A Context-Aware, Psychotherapeutic Music Recommender System for Commuters

Abstract

The advancements in urban commuting have enabled ease of travel for commuters. However, in the underdeveloped world, commuting has become a challenge for the mental health of commuters. A commuter who travels through public transport or their vehicle can develop depression and anxiety due to traffic congestion and unwanted delays. Symptoms of depression and anxiety can be mitigated through psychotherapeutic music. However, this music requires quiet rooms where a patient could listen to them. This can be overcome by playing music available on online streaming services via the commuters’ smart devices. The data from the sensors embedded in a commuter’s smart device is gathered and is termed the current context. The context includes both the data from the sensors and deduced data that is acquired through sensor services. The current context is then processed to determine the context of the commuter. The context is a label that is the outcome of a machine learning algorithm as part of context processing. The authors have utilized Bayesian probability to classify the current context of the commuter. Based on the classification outcome, which is termed context, a suitable playlist is generated and played on the commuters’ smart devices. A feedback loop enables improvement in classification as well as playlist generation. This proposed mechanism would improve the mental health of commuters including students, workers, and passengers, traveling to work and back frequently.