An approach which is based on joining acquisition of conceptual
graphs and analyzing them for further implementations is presented. The goal
of conceptual graphs analysis is to find informative conceptual graphs for
storing them in digital library. This analysis is based on learning links between
sentences in abstracts by using two instruments: conceptual graphs clustering
and passive voice grammar form as a possible pointer to the link between
sentences. All results of acquisition of conceptual graphs and their clustering
can be visually illustrated using modeling framework.