Temporal Knowledge Discovery with Infrequent Episodes

In this paper, we present an efficient algorithm which discovers rare episodes with a combination of bottomup and top-down scanning schema. The information sharing between bottom-up and top-down scannings helps prune candidate episodes, and thus, efficiently find infrequent episodes that are interesting to user. We evaluate the performance of the algorithm using real-life weather databases. We observe from experimental results that our approach results in 30%-90% reduction in computation time and 25%-75% reduction in the number of candidates comparing with Apriori algorithm.

• Type: Conference
• Year: 2004
• Where: DGO
• Authors: Dan Li, Liying Jiang, Jitender S. Deogun
• Keywords: DGO 2004, DGO 2007, Efficient Algorithm, Top-down Scanning Schema, Top-down Scannings