一种基于字典树的传感器节点关联规则的挖掘算法

A Lexicographic Tree Algorithm for Mining Association Rules from Wireless Sensor Networks

无线传感器网络中节点密集,分布范围广,长期监测使得信息量巨大,如何从大量的感知数据中提取或"挖掘"有用的知识,就成为无线传感器网络中信息处理的核心问题。本文提出一种新的关联规则挖掘算法PLT-STREAM,用来发现节点之间的有用关联,以此消除节点之间信息的冗余。该算法能帮助用户对数据进行有效的融合、分类、查询、分析、理解和决策。实验结果表明,该方法能够有效减少信息处理中通信和计算所消耗的能量,缩短数据查询响应的时间,从而延长整个网络的寿命。

Wireless sensor networks with high node density and wide node distribution, long-term monitoring produce a huge amount of data, so how to process the large data streams in sensor networks efficiently and find interesting knowledge in these streams become a new challenge. This paper proposes a novel node association rule mining algorithm PLT-STREAM for exploiting the inherent correlations between sensor readings. This algorithm can help users to manage data efficiently during aggregation, classification, prediction, query, understanding and decision-making. The experimental results show that the proposed method can reduce the overhead of computation and communication energy in the information processing procedure effectively. Our algorithm can also shorten the data query response time and thus prolong the network lifetime.