Stochastic blockmodeling of linked networks

Blockmodeling linked networks aims to simultaneously cluster two or more sets of units into clusters based on a network where ties are possible both between from the same set as well different sets. While this has already been developed for generalized and k-means blockmodeling, our approach is well-known stochastic blockmodeling technique, utilizing mixture model. Estimation performed using CEM algorithm, which iteratively estimates parameters by maximizing suitable likelihood function reclusters according parameters. The steps repeated until ceases improve. A key drawback basic algorithm that it treats all equally, consequently yielding higher influence larger parts data. greater size, however, does not necessarily imply importance. To mitigate asymmetry, we propose solution underrepresented data given through an appropriate weighting. This idea leads so-called weighted approach, ordinary replaced likelihood. efficiency approaches tested via simulations. It shown simulations performs better clearer blockmodel structure, especially when one-mode blockmodels within smaller clearer.