ISSN 1000-1239 CN 11-1777/TP

Journal of Computer Research and Development ›› 2017, Vol. 54 ›› Issue (1): 94-110.doi: 10.7544/issn1000-1239.2017.20150804

Previous Articles     Next Articles

The Maximal Regular Patterns Mining Algorithm Based on Landmark Window over Data Stream

Wen Yingyou1,3, Wang Shaopeng1,2, Zhao Hong1,3   

  1. 1(College of Computer Science and Engineering, Northeastern University, Shenyang 110819 ); 2(College of Computer Science, Inner Mongolia University, Huhhot 010021); 3(Key Laboratory of Medical Image Computing (Northeastern University), Ministry of Education, Shenyang 110819 )
  • Online:2017-01-01

Abstract: Mining regular pattern is an emerging area. To the best of our knowledge, no method has been proposed to mine the maximal regular patterns about data stream. In this paper, the problem of mining maximal regular patterns based on the landmark window over data stream is focused at the first time. In order to resolve the issue that the nave algorithm which is used to handle the maximal regular patterns mining based on the landmark window over data stream does not have the characteristic of incremental computation, the DSMRM-BLW(data stream maximal regular patterns mining based on boundary landmark window) algorithm is proposed. It takes the first window as the boundary landmark window, and handles it with the nave algorithm. For all other windows, it can obtain the maximal regular patterns over them based on the ones over the adjacent last window incrementally, and can overcome the drawback of the nave algorithm. It is revealed by the extensive experiments that the DSMRM-BLW algorithm is effective in dealing with the maximal regular patterns mining based on the landmark window over data stream, and outperforms the nave algorithm in execution time and space consumption.

Key words: data stream, landmark window, maximal regular pattern, incremental calculation, boundary landmark window technology

CLC Number: