Methods for graph-based representation of textual documents in natural language processing tasks

Akbasheva Eugeniya Amirovna, Akbasheva Galina Amirovna, Tlupov Islam Zaurbekovich

Nowadays, text is the most common form of information storage. Document representation is an important step in the process of data and text mining, natural language processing and information retrieval. The paper provides a formal description of the graph-based text representation problem, also discusses some natural language processing (NLP) problems which depend on different types of graph structures, advantages and disadvantages of methods used for graph-based text representation. In addition, a comparison of different approaches is given, as well as the applications in which these approaches are used.

Keywords:Graph, mathematical model, text analysis, natural language processing.