Bayesian networks to identify statistical dependencies. A case study of Spanish university students' habits.

OBJECTIVE: The present study sought to discover the relationships among different features characterizing Spanish university students' habits through a Bayesian network (BN). The set of features with the strongest influence in specific features can be determined.

METHODS: A BN was built from a dataset composed of 13 relevant features, determining the dependencies and conditional independencies from empirical data in a multivariate context. The structure was learned with the bnlearn package in R language introducing prior knowledge, and the parameters were obtained with Netica software. Three reasoning patterns were considered to make inferences: intercausal, evidential, and causal reasoning.

RESULTS: BN determined the different relationships. Through inference several conclusions were achieved, for instance a high probability value of physical activity in low state was obtained when active peers were instantiated to none state, self-rated fitness to fair state, bmi to normal weight, sitting time to moderate, age to 22-25, and gender to woman state.

CONCLUSIONS: Bayesian networks may help to characterize Spanish University students' habits.