A New Association Rule Mining Algorithm

A new algorithm called STAG (Stacked Graph) for association rule mining has been proposed in this paper using graph theoretic approach. A structure is built by scanning the database only once or at most twice that can be queried for varying levels of minimum support to find frequent item sets. Incremental growth is possible as and when new transactions are added to the database making it suitable for mining data streams. Transaction scanning is independent of the order of items in a transaction. Performance of this algorithm has been compared with other existing algorithms using popular datasets like the mushroom dataset, chess and connect dataset of the UCI data repository. The algorithm excels in performance when the dataset is dense.