Axiomatic Relation Extraction from Text in the Domain of Tourism.

Tourism is one of the most important activities in the economic sector. Thus, text about this topic coming from diverse data sources such as the Web should be analyzed to get information and knowledge that can be represented and consumed by people and applications. Such tasks are part of the Ontology Learning and Population (OLP) field, where the goal is to find elements of information (named entities) and their associations to create a knowledge base. In this regard, OLP deals with the discovery of named entities and how they are grouped, related, and subdivided according to their characteristics (class). However, the association between entities has not been studied in the same way, particularly axiomatic relations that exploit the type of the named entities. Therefore, this paper proposes a strategy for the extraction of axiomatic relations from text. It is based on the identification of named entities, their class, and their co-occurrence in the text to define lexical-syntactic patterns that support the extraction of axiomatic relations such as equivalence and disjointness between classes. The results demonstrate the usefulness of the strategy to produce new statements that enrich ontologies and knowledge bases.