An efficient one-pass method for discovering bases of recently frequent episodes over online data streams

摘要: The knowledge embedded in an online data stream is likely to change over time due the dynamic evolution of stream. Consequently, infrequent episode mining stream, frequent episodes should be adaptively extracted from recently generated segments instead whole However, almost all existing approaches find frequently occurring sequence. This paper proposes and investigates a new problem: streams. In order meet strict requirements such as one-scan, adaptive result update instant return, we choose novel frequency metric define highly condensed set called base episodes. We then introduce one-pass method for bases Experimental results show that proposed capable finding quickly adaptively. outperforms previous with advantages one-pass, more resulting sets less space usage.