Regresar

Hearing Health Virtual Assessment Through Association Rules Mining Inside a College Community

Abstract:

This study quantified possible cases of hearing loss and determined possible causes through association rules mining. The data was collected through an online survey that evaluated the medical background of the participants, and a free online hearing test provided by the American company Phonak. The study sample consisted of 226 entries, among students, faculty, and staff members of the Escuela Politécnica Nacional, a public university from Quito, Ecuador. For the initial data treatment and conversion to binary tables of the audiometry results, a Python optical character recognition algorithm was used. Finally, the association rules were obtained using the Apriori algorithm, which was implemented through the arules package for R. The results showed that 66.36% of the sample presents hearing loss in at least one ear. Furthermore, hearing loss causes in males are primarily related to prolonged use of headphones and loud noises, and these cases are mostly seen in younger individuals; while hearing loss causes in females are primarily related to family history of deafness.