Análisis comparativo de técnicas de minería de datos para el enriquecimiento de estructuras semánticas.

Abstract

This paper presents the comparative analysis of ontological mining techniques, describing the process of the algorithms. In the first place, the ontological mining is carried out semi-automatically and separately by leaving out all knowledge and obtaining weak semantic structures. Then, the comparative analysis is implemented on the techniques of alignment, linking and mixing of ontologies in order to integrate them. The first step is the alignment, then the linking and mixing of ontologies. The comparative analysis between two semantic structures is carried out into 3 ontological mining algorithms. First is to solve the semantic enrichment, then the alignment uses calculation of values of semantic similarities to obtain the strong final correspondences which take the name of ontology alignment. The linked it is done based on the alignment which represents all the equivalent concepts without being able to incorporate all the knowledge. The mixing also works on the alignment and managing in order to add all the knowledge left out in the link. The algorithms were tested in a web application, which helped achieved excellent results of being able to be compared and analyzed among them.