19th International CODATA Conference
Category: Knowledge Discovery
Peculiarity Oriented Mining For Heterogeneous Databases-A Multilayered Approach
R.Hemamaalini (saihemamalini@svce.ac.inm), Sri Venkateswara College of Engineering, India
Kanna
Nehemiah (chandrakhanna@annauniv.edu),
T.S.Nirmala (nirma@svce.ac.inm), Sri Venkateswara College of Engineering, India
Peculiarity oriented mining is one of the recent areas of active research. It
involves discovery of peculiarity rules by searching relevance among a set of
peculiar data. Peculiar data are those which are relatively small in number
and are different from the other data. Peculiarity mining can reveal novel and
interesting patterns unlike association mining and exception mining, which require
large support and hence may not reveal all interesting rules. Heterogeneous
databases are databases that have different structures, data types, stored formats,
or interpretations for the data. The semantic heterogeneity problem in such
databases can be solved by the multilayered approach of storing the data [1].
The existing approach [2] employs peculiarity mining in multidatabase.
Data mining can be used to view data at a high level of abstraction and low
level data can be transformed into high level data. Viewing data at a high level
may be less heterogeneous and can be easily communicated. In this paper framework
for peculiarity mining in such kind of heterogeneous databases is proposed.
The multilayered approach for mining in heterogeneous databases is combined
with the concepts of peculiarity mining to discover peculiarity rules in heterogeneous
databases. An attribute-oriented method of evaluating peculiarity factor to
discover peculiarity rules is also presented. Hence this approach can be efficient
in discovering interesting rules which can be viewed at different levels of
abstraction.
[1] Jiawei Han, Raymond T.Ng, Yongjian Fu,Son .Dao, “Dealing with Semantic Heterogeneity by Generalization-Based Data Mining Techniques” ACM
[2] Ning Zhong, Yiju Yao,Muneaki Ohshima, “Peculiarity Oriented Multidatabase Mining” IEEE Transactions on knowledge and data engineering, vol. 15, no. 4, July/August 2003.