Regresar

Machine learning applied to the analysis of alcohol consumption and its relationship with perceived stress

Abstract:

Alcohol consumption is one of the problems that affect the university population, particularly in their learning process and mental health. Previous studies carried out by the WHO and several scientists worldwide, link alcohol consumption with perceived stress. Machine learning techniques are being applied to various areas of science to establish analyzes from a complementary perspective to conventional statistics, presenting relevant results for the development of these sciences. In the Project “CEPRA XII-2018-05 Pbkp_rediction of Drug Use”, 7134 clean records were obtained from the Alcohol Use Disorders Identification TestSelf-Report Version (AUDIT-C) and Perceived stress scale (PSS-10), together to sociodemographic data of Ecuadorian university students. Using the Lanz (2013) methodology, the k-means and hierarchical algorithms were applied, obtaining three groupings related to alcohol consumption and stress, which corroborate the theory.