Semi quantitative classification of relative importance of concepts in text, ranks concepts according to their importance from ontology and user preferences

摘要

Information is characterized by the concepts described in the text and these concepts are ranked according to their importance for characterizing the information in the text in which they are found, using a method comprising the following steps: (1) determining the intrinsic importance of a concept from an ontology and according to a user preference; (2) determining the statistical, semantic, position-determined and typographical-determined importance of the concepts in the text; (3) aggregating these importance values to provide a preliminary measure of the importance of all the concepts; (4) determining the importance of all concepts from the ontology according to the structural relations in the ontology, on the basis of the results of the previous step; (5) aggregating the importance values from the previous two steps to provide another preliminary measure of the importance of all the concepts; (6) determining the statistical relation-determined and semantic relation-determined importance of the concepts in the text; (7) aggregating the importance values determined in the two previous steps to provide another preliminary measure of the importance of all the concepts; and (8) repeating the previous four steps until there is no more significant difference in the result of the previous step, providing a final measure of the importance of all the concepts.