Predicción del consumo de drogas y juegos de azar en jóvenes deportistas a través de un modelo de RNA

Introduction: adolescence is established as the critical stage for the onset of consumption of addictive substances. The aim of this study is to identify the pattern of lifetime consumption of the main drugs at this stage, analysing various socio-contextual variables. Objective: identify the level of drug and gambling consumption in life by means of socio-demographic variables. Methodology: this study uses ex post facto correlational-predictive design with 256 football and futsal players from the capital of Jaén. Information was obtained on the following socio-demographic variables: father's and mother's employment status, father's and mother's academic level, family economic level and weekly money spent on leisure. Results: the construction of the ANN model showed an overall correct percentage in training of 83.4% and in the test of 77.6%. The most important predictors of lifetime use of alcohol, tobacco, cannabis and gambling are weekly leisure money (100%) and family finances (57.2%), while the least important predictor is the mother's employment status (23%). The ROC levels show moderate-good values, highlighting the prediction of consumption values above 39 days with a mean of around 0.872. Conclussion: lifetime use of the addictive substances analysed cannot be approached in a linear fashion. It is necessary to know the socio-demographic factors of the adolescents and the interaction between them to be able to identify possible risk subjects.