Fig. 2

Defining an abstraction level based on information content (IC) of terms. Blue boxes denote terms that are considered under the given abstraction layer (here at IC range from 4.0 to 7.5); green circles denote terms that are used for representing these term aggregations, crossed out terms are ignored. Values in blue circles denote the information content (IC) of the term. Examples: Left box: Aneurysm, Ruptured and Aneurysm are aggregated into and represented by Vascular Diseases. Center box: Cardiovascular Abnormalities and Congenital Abnormalities are considered but Cardiovascular Diseases is ignored (too general). Demyelinating Diseases will not be considered under this abstraction layer because there is no ancestor term into which they could be aggregated. Tree numbers in square brackets indicate hierarchical relations between terms according to the MeSH tree hierarchy