1999 | OriginalPaper | Chapter
Discovering and Visualizing Attribute Associations Using Bayesian Networks and Their Use in KDD
Authors : Gou Masuda, Rei Yano, Norihiro Sakamoto, Kazuo Ushijima
Published in: Principles of Data Mining and Knowledge Discovery
Publisher: Springer Berlin Heidelberg
Included in: Professional Book Archive
Activate our intelligent search to find suitable subject content or patents.
Select sections of text to find matching patents with Artificial Intelligence. powered by
Select sections of text to find additional relevant content using AI-assisted search. powered by
In this paper we describe a way to discover attribute associations and a way to present them to users using Bayesian networks. We describe a three-dimensional visualization to present them effectively to users. Furthermore we discuss two applications of attribute associations to the KDD process. One application involves using them to support feature selection. The result of our experiment shows that feature selection using visualized attribute associations works well in 17 data sets out of the 24 that were used. The other application uses them to support the selection of data mining methods. We discuss the possibility of using attribute associations to help in deciding if a given data set is suited to learning decision trees. We found 3 types of structural characteristics in Bayesian networks obtained from the data. The characteristics have strong relevance to the results of learning decision trees.