Models of representation of social mobility and inequality systems. A neural network approach

This paper reports on the results of the application of an innovative technique, i.e. neural network models, to mobility data. Our primary aim is to show that the technique is more flexible than traditional statistical modeling, and that it entails less strong methodological assumptions concerning the phenomenon which they are intended to represent. Two kinds of networks have been applied: heteroassociative networks, used for prevision and class membership recognition; and autoassociative networks, used for simulation tasks. Results obtained from experiments with neural networks on Italian data are highly consistent with the body of knowledge derived from previous classical analysis. The explicative power of neural network models proved to be higher than that of path analysis given their capacity to uncover any kind or relation between variables, whether linear or nonlinear. When compared to log-linear models, they enable the reconstruction of mobility processes within a global frame, controlling all relevant variables at once.