Lecture Notes in Computer Science, 2000, Volume 1967/2000, 285-289, DOI: 10.1007/3-540-44418-1_31

Knowledge Integration of Rule Mining and Schema Discovering

Kohei Maruyama and Kuniaki Uehara

View Related Documents

Abstract

Despite the growing popularity of semi-structured data such asWeb documents and bibliography data, most data mining researches have focused on databases containing well structured data like RDB or OODB. In this paper, we try to find useful association rules from semi-structured data. However, some aspects of semi-structured data are not appropriate for data mining tasks.
One problem is that semi-structured data contains some degree of irregularity and it does not have fixed schema known in advance. The lack of external schema information make it a very challenging task to use standard database access method or to apply the algorithms of rule mining. Therefore, schema discovering is considered to be necessary for rule mining.
Another problem of association rule mining is computing cost. If discovered schema pattern contains redundant attributes, they affect mining efficiency. Therefore, we try to feedback knowledge that obtained from the result of association rules to schema discovering. It means rule mining and schema discovering can give benefit to each other. In this way, by integrating knowledge of both rule mining and schema discovering, we can extract useful association rules from semi-structured data efficiently.

Fulltext Preview

Image of the first page of the fulltext document