期刊名称：International Journal of Electronics Communication and Computer Engineering
印刷版ISSN：2249-071X
电子版ISSN：2278-4209
出版年度：2012
卷号：3
期号：5
页码：1270-1278
出版社：IJECCE
摘要：Data mining techniques with the existing approaches look for data as patterns in a single data table. In recent years, the most common types of patterns and data mining approaches are extended to the multi-relational data. Multi-Relational Data Mining (MRDM) approaches involve multiple tables (relations) from a relational database. MRDM now encompasses association rule discovery, decision trees and distance-based methods among others. Three popular techniques such as classification, clustering and association rules have been explored in propositional and structured data mining. It then provides an insight into various classification methods including Inductive Logic Programming, Relational database, emerging patterns and associative approaches. MRDM approaches have been successfully applied to a number of problems in a variety of areas especially in the area of bioinformatics. This paper discusses some of the characteristics and comparisons in detail and finally on future directions of some applications of MRDM.
关键词：Classification; decision tree; inductive logic programmig; multi relational data mining
