Authors:
F. H. Ismail
1
;
H. K. Mohammed
1
;
M. A. Ismail
2
and
I. EL-Maddah
2
Affiliations:
1
Faculty of Computer Science, Misr International Univ. Cairo; Faculty of Engineering, Ain Shams Univ. Cairo, Egypt
;
2
Faculty of Engineering, Ain Shams Univ. Cairo, Egypt
Keyword(s):
Mining association rules from semi-structured data, meta-rule guided mining of association rules, discovering frequent structure from XML.
Related
Ontology
Subjects/Areas/Topics:
Artificial Intelligence
;
Knowledge Discovery and Information Retrieval
;
Knowledge-Based Systems
;
Soft Computing
;
Symbolic Systems
;
Web Mining
Abstract:
Nowadays, some information is semi-structured. The main characteristic of semi-structured data (XML) is
that they have irregular structure. There is no distinction between data and structure. Even though, it is quite
common that semi-structured objects representing the same sort of information have similar, though not
identical, structure (pattern). Previous work has introduced templates for mining association rules from
XML based on prior knowledge about the structure of the XML document. If the users do not have any
knowledge about the structure in advance, what would be their clue in writing templates? In this paper, we
introduce a new approach for designing association rule templates based on the automatic discovery of
frequent structure in the XML document. Frequent structure serves as a schema built over the semi-structured
data. This layer guides the user to the useful structure that might yield useful associations rather
than choosing any piece of structure at random
. The structured layer is displayed from which the user can
select templates of interest. Association rules that comply with the specified templates are generated.
(More)