Análisis multiobjetivo y modelos de regresión. Una aplicación para analizar el bienestar de los estudiantes españoles

Abstract

In this work, a novel approach is proposed in which econometric and multiobjective optimization techniques are combined with the aim of analysing socio-economic problems. This approach consists of two stages. Firstly, a regression model is carried out, from which, in the second stage, a multiobjective optimization problem is defined. An advantage of this combination is the possibility of introducing preferences of decision makers -desired values- for solving the problem. Particularly, we apply this approach to analyse the well-being of Spanish students through four indexes (positive feelings, motivation, sense of belonging, and bullying). In recent years, studying the students’ well-being has become very relevant because of its relation with their academic performance. Thus, four regressions are obtained (one per index) with respect to a set of explanatory variables, from which the multiobjective optimization problem is built. The results obtained using interval multi-objective programming provide us information both about how the improvement of one index can affect the values of the remaining ones, and also about the student’s profile who achieves an optimum balance among the well-being indexes.