Profile association rule mining from mixed database

Association rule mining from a transactional database or from a relational database takes a great deal of attention from the researcher. Various algorithms have been developed to extract association rules from either relational or transactional databases. A need exists to develop a methodology to extract association rules from combined (or mixed) relational and transactional databases. In this paper, a new algorithm definition introduced to extract association rules from transactional and relational databases or a mixed database. The developed algorithm extracts a profile association rules in the form: IF customer profiles THEN customer behavior. The new algorithm is applicable in both industrial and academic settings, where finding a relationship between the customer and his/her behavior is needed. Performance testing indicates that the new algorithm results in an improvement over earlier work and can be improved further in the future.