SAFARM: simulated annealing based framework for association rule mining
摘要
The research paper introduces an algorithm called SAFARM which performs association rule mining with the help of simulated annealing. It’s a multi-objective problem with vast search space. The suggested approach is independent of the database as it does not require minimum support or minimum confidence specification. In the algorithm, a fitness function is designed to fulfill the required objective and the presentation of rules is proposed with a compact structure. The correctness and efficiency of the algorithm is verified by testing it on synthetic and real databases.