Mining of association rules in a relational database is important because it discovers new knowledge in the form of association rules among attribute values. This enables business managers to make the right decisions pertaining to their businesses. However, association rule mining concepts and algorithms have been traditionally applied to a specific representation of data called market basket data representation or transactional representation. Data in a relational database is not represented in a market basket data format but rather in a format that adheres to the relational data model. Data in a relational database has to be converted to the market basket data representation before data mining algorithms can be applied to it. This requires the application of tedious conversion processes on large quantities of data before such algorithms can be applied. In this paper, we describe how association rules can be expressed directly in the context of the relational data model, and not based on market basket data representation. Many of the association rule mining concepts are re-defined in this paper in a way that makes them conform to the relational model of data.
Data Mining, Association Rule, Itemset, Relational Model, Relational Database