DSpace

Future University Hakodate Academic Archive >
研究者 >
情報アーキテクチャ学科 >
高橋 修  >

このアイテムの引用には次の識別子を使用してください: http://hdl.handle.net/10445/5998

タイトル: A Method for Sensor Data Prediction Based on Correlations among Multiple Data Streams
著者: Shiraishi, Yoh
Arai, Kenji
Takahashi, Osamu
アブストラクト: By developing sensor network technologies, researches on data stream processing and stream mining have received much attention. Stream data prediction which predicts future data pattern by analyzing tendency of past data stream is one of streammining techniques. In a sensor network deployed in the real-world such as a city, a building and a room, data stream from a certain sensor node may relate to that from other ones. It is expected to improve the accuracy on sensor data prediction by considering correlations of these sensor data streams. In this paper, we propose a method for sensor data prediction based on correlations of multiple sensor data streams. This method extracts the feature quantities from partial sequences of each data stream and classifies these sequences into multiple groups by a clustering algorithm. The classified group is a cluster that expresses a pattern from the stream. Our method uses correlations among these clusters from different kinds of data streams and correlations between the past sequences on a stream in order to predict future data pattern. This paper describes an overview of our method for sensor data prediction and reports the results of the preliminary experiment. 
研究業績種別: 国際会議/International Conference
資料種別: Conference Paper
査読有無: あり/yes
単著共著: 共著/joint
開始ページ: 25
終了ページ: 30
年月日: 2010年10月
出版社: IEEE
出現コレクション:高橋 修

ファイルダウンロード:

このコンテンツにファイルはありません。

このアーカイブに登録されているコンテンツはすべて著作権により保護されています。
著作権を遵守の上、ご利用ください。

 

Copyright © 2010-2012 FUTURE UNIVERSITY HAKODATE.
Powered by DSpace Software Copyright © 2002-2007 MIT and Hewlett-Packard