A Biclustering Approach via Mixture of Latent Trait Analyzers for the Analysis of Digital Divide in Italy

We employ an extension of the Mixture of Latent Trait Analyzers (MLTA) model to analyse the digital divide in Italy in a biclustering perspective. In detail, units (individuals) are partitioned into clusters (components) via a finite mixture of latent trait models; in each component, variables (digital skills) are partitioned into clusters (segments) by modifying the linear predictor's specification of the original MLTA model. This allows us to identify homogeneous groups of individuals with respect to subsets of digital skills, also accounting for the influence of demographic features on the probability of being digitally skilled.