Sugestões de relações entre elementos de uma ontologia através de análises de textos.

Abstract

Today we are faced with large amounts of information throughout the world we live in and especially when we are connected to the web. These data alone cannot be very significant, but if they are analyzed and related to each other, they can reveal more information than they appear initially, thus enabling the creation of a large network of relationships. The work proposed here is to implement a module for OntoClipping software, which has an ontology formed from electronic clipping held in web pages, originally developed by Claudio Omar Correa Carvalho Junior. This module will aim to suggest relationships for individuals who are mentioned in the ontology. For example, we imagine an ontology where we have individuals John and Peter, who belong to the Person class, and also individuals Legend and Sunset belonging to the Party class. Both John and Peter, have attended these festivals, so we have a relationship between these individuals. To create these relationships will need to perform the analysis of texts that are in OntoClipping database and, on top of that, develop ideas and algorithms for solving the problem.