In this talk, we present two algorithms for determining the relevance of process information in POIL. The first one determines the popularity of process information based on the relationships of a SIN. The second one determines the popularity of process information based on user ratings. We applied our algorithms to a real-world scenario, i.e., validated them based on an implementation and a survey in the automotive domain.
Have a look on the related paper: 
http://dbis.eprints.uni-ulm.de/955/