A Method for Heterogeneous Distributed Data Streams Classification Based on Sprint
Currently,the application of heterogeneous distributed data streams is more and more pervasive (such as sensor networks,process control,etc.).So,the capacity of extracting knowledge from Heterogeneous distributed data streams has become very important However,the classification research of heterogeneous distributed data streams has just started.This paper presents a method for classifying heterogeneous distributed data streams based on sprint.It adopts a distributed stream processing architecture;Local nodes calculate split points and perform a split in parallel,then send to the central site.The central site determines the final split of the current node.It only transfers a small amount of information for decision-making,not to transmit raw data between local sites and the central site,thus effectively reducing the traffic load.
data mining data stream distributed data stream heterogeneous distributed data stream
Huafcn Xu Xingjun Ping
Department of Computer Science North China Institute of Science and Technology Hebei,China Huangtuzhuang Local Police Station Public Security Bureau Hebei,China
国际会议
太原
英文
183-187
2011-02-26（万方平台首次上网日期，不代表论文的发表时间）