Different approaches for miningassociation rules from Multi relational databases

Abstract

Defining a multi relational pattern that is suitably meaningful and discovering vital information through efficient mining algorithm is a challenging problem and significant works have been presented in the literature to solve this challenging task Accordingly an efficient approach for effectual mining of relational patterns is developed from the multi relational database Primarily the multi relational database is represented using a tree based data structure without changing the relations A tree pattern mining algorithm is designed and applied on the constructed tree based data structure for extracting the multi relational frequent patterns The designed mining algorithm makes use of positional data information from the tree data structure so that the required time for mining algorithm is more efficient and effective Then in the second phase of the research the adaptation of GA is done to optimize the multi relational pattern to rule mining The Genetic algorithm Optimization algorithm together with Mining algorithm helps to mine the most significant interesting rules An optimized rule generation with the newly designed fitness function of genetic algorithm is developed Initially every rule is evaluated based on the fitness function and GA has sorted the evaluated rules based on the fitness values subsequently it selects the rules which are greater than the fitness rate as optimal rules newline