Author(s)

E. Bezerra, M. Mattoso & G. Xexeo

Abstract

In this paper we present a classification of integration frameworks found in the literature for the coupling of a data mining application and a DBMS. We classify the database coupling in several categories. Along these categories, we analyse several issues in the integration process such as degree of coupling, flexibility, portability, communication overhead, and use of parallelism. We also present the trade-off of using each one of the integration frameworks and describe the situations where one framework is better than the other. We also describe how to implement DBMS integration in several Data Mining methods and we discuss implementation aspects including parallelism issues for each one. We analyse these solutions and show their advantages and disadvantages. 1 Introd

Keywords