Ederson Garcia,  Marina Teresa Pires Vieira. 

This paper presents a case study of multi-relational data mining using the ConnectionBlock algorithm, applied to the database of a sugar mill. The algorithm handles multiple tables not explicitly correlated but which influence one another according to the semantics of the data involved. The experiment revealed very interesting and useful patterns that are not found using traditional algorithms. The paper aims to present how the data were prepared to obtain better expressiveness of the rules generated, showing the potential of the algorithm to find patterns in semantically related data.

http://www.lbd.dcc.ufmg.br:8080/colecoes/sbbd/2008/016.pdf

Processando consulta.