Full metadata record
DC FieldValueLanguage
dc.contributorDepartment of Computingen_US
dc.creatorDing, Xiaoping-
dc.identifier.urihttps://theses.lib.polyu.edu.hk/handle/200/1843-
dc.languageEnglishen_US
dc.publisherHong Kong Polytechnic University-
dc.rightsAll rights reserveden_US
dc.titleMining wireless sensor network data streams using distributed algorithms and granularity algorithms frameworken_US
dcterms.abstractWireless Sensor Networks (WSN) is networks of wireless sensor nodes, sensory devices with limited energy, and a limited capacity to compute or communicate wirelessly. Data Streams Mining (DSM) is the process of extracting knowledge structures represented in models and patterns in continuous streams of information. DSM programs usually require a mass of computing and communication. Typically, all data streams in all nodes on WSN are collected and clustered at one center node using traditional centralized algorithms such as K-Means Clustering. However, such centralized algorithms require large volumes of inter-node communication and this shortens the life of a WSN. In this dissertation, we propose Distributed K-Means Clustering (DKMC), which distributes the average task of clustering computing to every node in a WSN, and which collects the results in every node to K central nodes for averaging energy and addressing constraints from WSN. As the basis of the DKMC algorithm, we developed three algorithms, Count, Sum, and Average. We also present the Difference-Value monitoring, a novel and more stable type of data stream monitoring that addresses the fact that different sensory devices have different error ranges. This paper also seeks to make better use of limited WSN node resources by adjusting the granularity, or precision, of data at each stage of data mining process. Previous, [MAS2004] resource aware, approaches have improved the efficiency of resources use by controlling the granularity of just one phase of DSM, the output phase. In this work, we propose a Granularity Algorithm Framework (GAF). The GAF saves energy by adjusting data granularities at WSN nodes in three phases: the data stream input phase, the data streams mining phase, and the data stream output phase. The GAF adjusts granularities using three parameters: task (e.g. precision of mining result), environment (e.g. bandwidth and network topology) and resource ( e.g. CPU, memory). This is a significant improvement on the resource aware approach which adjusts granularity according to only one parameter, resources.en_US
dcterms.extentix, 112 leaves : ill. ; 30 cm.en_US
dcterms.isPartOfPolyU Electronic Thesesen_US
dcterms.issued2007en_US
dcterms.educationalLevelAll Masteren_US
dcterms.educationalLevelM.Sc.en_US
dcterms.LCSHHong Kong Polytechnic University -- Dissertations.en_US
dcterms.LCSHSensor networks.en_US
dcterms.LCSHWireless LANs.en_US
dcterms.LCSHData mining.en_US
dcterms.accessRightsrestricted accessen_US

Files in This Item:
File Description SizeFormat 
b21288719.pdfFor All Users (off-campus access for PolyU Staff & Students only)17.68 MBAdobe PDFView/Open


Copyright Undertaking

As a bona fide Library user, I declare that:

  1. I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
  2. I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
  3. I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.

By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.

Show simple item record

Please use this identifier to cite or link to this item: https://theses.lib.polyu.edu.hk/handle/200/1843