Rule discovery on related objects

Abstract- Many-to-many relations are often observed between real life objects. When many-to-many relations are between objects in the same class the data mining process becomes more complicated than mining objects when there are no such recursive relations. Mining objects re-lated to other objects in the same class requires construction and execution of recursive queries and hence interpretation of the results of recursive queries. Here, we describe Rila, a new rela-tional rule discovery system that was extended for mining objects related to other objects in the same class. Experimental results are provided on Mutagenesis and KDD Cup 2001 data sets. Current relational database systems utilize many advanced technologies to store and serve structured data, such as indexing and query services and transaction management support. For this reason, a significant amount of today’s complex scientific and commercial data is stored in relational database systems