An approach on association patterns mining in telecommunication alarm database

Currently limiting to the minimal support, the algorithms used in alarm association rules mining are almost applied in the frequently occurring alarm events. A new algorithm based on graph theory to mine telecommunication alarm pattern is proposed. It first mines network's 2-items association pattern by scanning the database only once, and then gets the maximal association mode, so that it can avoid generating lots of middle items. Experiments based on the actual network alarm data demonstrate the efficiency and the effectiveness of the algorithm.