PRELIMINARY CLUSTERING ANALYSIS OF NOISE IN A HOSPITAL ROOM

In recent years, noise levels in hospitals and healthcare facilities have been on the rise. This becomes a significant concern as high levels of noise can negatively affect patients' recovery and interfere with communication between healthcare providers and patients. Furthermore, hospital noise can also have a negative impact on staff, leading to increased stress, fatigue and decreased job satisfaction. The present paper investigates environmental noise in a hospital setting by analyzing the sound pressure levels and comparing it with the WHO guidelines' limitations. Moreover, this work tries to overcome the classical analysis by exploiting the Gaussian Mixture Model (GMM) clustering algorithm to separate the noise sources during the morning round into two different clusters: mechanical and human noise. The mechanical noise cluster is further analyzed by comparing it with the sound pressure level collected in another room where only the mechanical sources were on. The results show that the GMM can provide important details of noise sources even in occupied conditions and that the proposed approach can effectively identify different noise sources in a hospital environment, which can potentially help in the development of noise control strategies and improve the overall acoustic environment for patients and staff.