Real Time Distributed Data Mining for Sensor Networks

传感器网络的实时分布式数据挖掘

• 批准号: 0239914
• 负责人: Sitharama Iyengar
• 金额: $ 3.27万
• 依托单位: Louisiana State University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2002
• 资助国家: 美国
• 起止时间: 2002-09-15 至 2003-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=0239914&HistoricalAwards=false
• 关键词: Real; Time; Distributed; Data; Mining

Data mining is the semi-automatic discovery of patterns, associations, changes, anomalies, rules, and statistically significant structures and events in data. Thus, data mining aims to extract knowledge from sensor data, and is an emerging area with numerous commercial applications. Work in data mining ranges from theoretical work on the principles of learning and mathematical representations of data to building advanced engineering systems that perform information filtering on the web, find genes in DNA sequences, help understand trends and anomalies in economics and medicine, and detect network intrusion. The problem gets more complicated in a distributed system where the data itself is positioned across different sites and coordination across the sites needed to mine information from the distributed data. Distributed Data Mining (DDM) arises in diverse areas such as sensor stream data generated from satellite etc. Presently, there are several proposed ad-hoc approaches to distributed data mining (DDM) that deal with specific schemes and most of them do not address any real time analysis for solving different classes of distributed data mining problems. This project will address this very explicitly the merits of the new framework.

数据挖掘是半自动地发现数据中的模式、关联、变化、异常、规则以及统计上重要的结构和事件。因此，数据挖掘旨在从传感器数据中提取知识，并且是具有许多商业应用的新兴领域。数据挖掘的工作范围从学习原理和数据数学表示的理论工作到构建先进的工程系统，这些系统在网络上执行信息过滤，在DNA序列中找到基因，帮助理解经济和医学的趋势和异常，以及检测网络入侵。 这个问题在分布式系统中变得更加复杂，其中数据本身位于不同的站点，并且需要从分布式数据中挖掘信息。分布式数据挖掘（DDM）出现在不同的领域，如从卫星等产生的传感器流数据，目前，有几个建议的ad-hoc分布式数据挖掘（DDM）的方法，处理特定的计划，他们中的大多数不解决任何真实的时间分析解决不同类别的分布式数据挖掘问题。 本项目将非常明确地解决这一问题，体现新框架的优点。