Building and Querying Large Modelbases

Abstract

Model building is one of the most important objectives ofdata mining and data analysis. As many data miningapplications, such as personalization, bioinformatics andsome large enterprise-wide business applications, becomeincreasingly complex and require a very large number ofmodels, it is becoming progressively more difficult for dataanalysts to built and to manage a large number of modelsin these applications on their own. Therefore, developmentof software tools helping data analysts in these tasks isbecoming a pressing issue. This paper presents a modelmanagement system supporting various types of datamining models. It describes how to build and populatelarge heterogeneous modelbases. It also presents a querylanguage for querying these modelbases and examinesperformance results for some of the queries.