Automatic Construction of Concept Maps from Texts

The automatic construction of concept maps from texts is still an ongoing research topic, especially when the map should represent, by means of concepts and relationships, a summarization of a complex text. We propose a new method for the automatic generation of concept maps. Based on an analysis of the limitations and best features of the state-of-the-art, this paper introduces a pipeline that comprises (a) grammar rules and depth-first search for the extraction of concepts and links between them from text; (b) anaphora resolution; (c) exploitation of named entities for concept identification; and (d) concept relevance calculation through frequency of occurrence and an analysis of the topology of the concept map. The approach and preliminary results are presented in the form of a prototype, which is still under development.