Authors:
Bassam M. Aoun
and
Marie Khair
Affiliation:
Notre Dame University, Lebanon
Keyword(s):
Semantic Web, Ontology, Web Mining, Text Mining, Apriori Algorithm.
Related
Ontology
Subjects/Areas/Topics:
Cloud Computing
;
Enterprise Information Systems
;
Semantic Web Technologies
;
Services Science
;
Software Agents and Internet Computing
Abstract:
The Semantic Web is a network of information linked up in such a way as to be easily processed by machines, on a global scale. To reach semantic web, current web resources should be automatically translated into semantic web resources. This is usually performed through semantic web mining, which aims at combining the two fast-developing research areas, the Semantic Web and Web Mining. A major step to be performed is the ontology-learning phase, where rules are mined from unstructured text and used later on to fill the ontology. Making sure that all rules are found and no additional and inaccurate rules are inserted, remains a critical issue since it constitutes the basis for building the semantic web. The mostly used algorithm for this task is the Apriori algorithm, which is inherited from classical data mining. However, due to the nature of the semantic web, some important rules can be dropped. This paper presents an enhanced version of the Apriori algorithm, En_Apriori, which uses
the Apriori algorithm in combination with the maximal association and the X2 test to generate association rules from web/textual documents. This provides a major refinement to the classical ontology learning approach.
(More)