Synthesizing Global Negative Association Rules in Multi-Database Mining

Association rule mining has been widely adopted by data mining community for discovering relationship among item-sets that co-occur together frequently. Besides positive association rules, negative association rule mining, which find out negation relationships of frequent item-sets are also important. The importance of negative association rule mining is accounted in customer-driven domains such as market basket analysis for identifying products that conflict with each other. In multi-database mining context, mining negation relation among item-sets and synthesizing global negative association rules from multiple data sources located in different places are having importance in arriving decisions both at strategic and branch levels. This paper made an attempt for synthesizing global negative association rules which are voted by most of the participating data sources while mining multiple data sources. Experimental data are employed to test the theoretical analysis of the proposal using UCI machine learning repository data set. The space and time complexity analysis presented in the paper show the efficiency of the proposed approach