Abstract
With the advances of wireless sensor networks, they yield massive volumes of disparate, dynamic and geographically-distributed and heterogeneous data. The data mining community has attempted to extract knowledge from the huge amount of data that they generate. However, previous mining work in WSNs has focused on supporting simple relational data structures, like one table per network, while there is a need for more complex data structures. This deficiency motivates XML, which is the current de facto format for the data exchange and modeling of a wide variety of data sources over the web, to be used in WSNs in order to encourage the interchangeability of heterogeneous types of sensors and systems. However, mining XML data for WSNs has two challenging issues: one is the endless data flow; and the other is the complex tree structure. In this paper, we present several new definitions and techniques related to association rule mining over XML data streams in WSNs. To the best of our knowledge, this work provides the first approach to mining XML stream data that generates frequent tree items without any redundancy.
Highlights
Wireless sensor networks (WSNs) have been identified as an important research area for the 21st century [1]
Throughout the paper, we focus on embedded subtrees from the dataset of XML stream data and use them in providing the definitions for association rules
Since we focus on the rule detection from XML stream data, each XML document corresponds to a set of XML data in a sink node, and the data stream is a continuous sequence of XML data blocks
Summary
Wireless sensor networks (WSNs) have been identified as an important research area for the 21st century [1]. The frequency of errors in the outputs generated by the online algorithms should be constricted to be as small as possible Due to these differences, previous multiple-pass data mining techniques presented for traditional data sets cannot be directly applied to the domain of mining the stream data. To the best of our knowledge, our proposed scheme is the first approach to mining XML stream data in the sense that it generates frequent tree items without any redundancy (see Section 4 for the definition of a tree item). The rest of this paper is organized as follows: Section 2 discusses prior work related to mining association rules from sensor data and XML data.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
