文献类型：期刊文献

英文题名：Sliding-window Top-k pattern mining on uncertain streams

作者：Zhang, Xiaojian[1]; Zhang, Yadong[1]

第一作者：张啸剑

通讯作者：Zhang, X.

机构：[1] School of Computer Science and Information Engineering, Henan University of Economics and Law, 450002, China

第一机构：河南财经政法大学计算机与信息工程学院

年份：2011

卷号：7

期号：3

起止页码：984-992

外文期刊名：Journal of Computational Information Systems

收录：EI(收录号：20111513904951);Scopus(收录号：2-s2.0-79953743339)

语种：英文

外文关键词：Algorithms - Information management - Network management - Sensor networks

摘要：Uncertainty pervades many application fields such as sensor network and mobile data management. In these fields, uncertain data items often arrive rapidly and need to be handled in a streaming fashion. The key challenge of processing uncertain streams stems from the limited memory and the CPU resource of handling both arriving and expiring windows in the high-rate streams, combined with the difficulty of coping with the dilemma caused by error parameter Ε. Setting Ε too high may obtain inaccurate results while setting it too low will make memory consumption large. In order to deal with the challenges, this paper focuses on finding the Top-k Patterns on uncertain streams, and based on the sliding window model and Chernoff Bound technology proposes a space- and time-efficient algorithm, called Topk-PU, in which an increasing expected support function is designed to approximately calculate the count of each pattern. Experimental results show that the fast processing rate, and the efficiency of our proposed algorithm. Copyright ? 2011 Binary Information Press.