Desarrollo de modelos predictivos basados en estudio epidemiológico de consumo de drogas en Universidades Ecuatorianas.

Abstract

Abstract: Drug s consumption is a social problem of urgent attention, with consequences that affect not only users but also the community in general, with young people being the most likely to fall into this type of behavior. Previous research carried out by the WHO and the scientific community at a global level, relates alcohol consumption directly with perceived stress and it s been considered a possible indirect relation with other factors such as loneliness, depression, psychological inflexibility and cognitive impulsivity. This last part focuses the effort of this study, applying Machine Learning techniques, which have been entering various branches of science to offer analysis from a complementary perspective to traditional statistics, generating relevant results for the development of these sciences. Using information regarding substances use and psychosocial factors from Ecuadorian university students; clustering algorithms and neural networks were applied, obtaining a significant value in the evaluation metrics.