Detecting and monitoring abrupt emergences and submergences of episodes over data streams

摘要: Existing studies on episode mining mainly concentrate the discovery of (global) frequent episodes in sequences. However, are not suited for data streams because they do capture dynamic nature streams. This paper focuses detecting changes frequencies over time-evolving We propose an efficient method online detection abrupt emerging and submerging Experimental results synthetic show that proposed can effectively detect defined patterns meet strict requirements stream processing, such as one-pass, real-time update return results, plus limited time space consumption. real demonstrate detected by our natural meaningful. The has wide applications monitoring analysis discovered indicate emergences/disappearances noteworthy events/phenomena hidden