Automatic Merging of Knowledge using Ontologies
Alma Delia Cuevas Rasgado, Adolfo Guzmán Arenas · Research in Computing Science · 2006
The fact that many people simultaneously construct the pages of the Web in an independent way, generates a great obstacle for the machines that track the information in it. Therefore, the concept of Semantic Network has been introduced. It provides a standardization of the information through markup languages (SGML, XML, etc.) where the user generates his own annotations, almost all of them as labels or syntactic rules. Relatively few of the languages have try to represent and to manipulate the knowledge with methods of Artificial Intelligence. This paper proposes a structure (an ontology) more suitable to represent the knowledge, with interesting contributions with respect to current languages (AML+OIL[5], RDF[8], OWL[12]). Also, this paper presents an automatic algorithm to match and merge two or more ontologies. This merging is important when it is desired to increase the knowledge in an ontology. In that way it is possible to accumulate the knowledge in an automatic way. The process of merging begins by obtaining the value of the similarity between each elements of the ontologies (through COM[1] Algorithm); later, the optimal matching is sought. Finally, the result defines the new ontology. This process is performed totally by the computer. That is to say, the user does not take part in this process, as it happens in current merging algorithms (OntoMerge[6], FCA-Merge[9], Chimaera[11], Prompt[13], If-Map[14]). In the merging, the OM Algorithm solves problems of contradiction and reorganization of the final ontology. The efficiency of the algorithm of fusion is demonstrated through several examples.