A Dynamic Solution for Association Rules Mining

We consider the problem of discovering association rules between items in a large database of sales transactions. We present a new algorithm for solving the problem that is fundamentally different from the known algorithms. This algorithm solves most of the major problems of well known association rules mining algorithms like Apriori and FP-tree. In this algorithm we generated mutually independent candidates as per our requirement in a single scan. It is very flexible to use in interactive mining system. During the interactive mining process, users may change the threshold of support according to the rules. In our algorithm once candidates are generated we can change the support without repeating the whole process. It is suitable for incremental mining. Since as time goes on databases keep changing, new datasets may be inserted into the database. In our algorithm insertion of these datasets is possible without repeating of the whole process. We can generate the candidates for required items and required itemsets in a single scan. This makes effective memory utilization and improves the performance.