Detecting Abnormal Trend Evolution over Multiple Data Streams

Authors:
Chen Zhang;Nianlong Weng;Jianlong Chang;Aoying Zhou
Affiliations:
Department of Computer Science and Engineering, Fudan University, P.R.C;Department of Computer Science and Engineering, Fudan University, P.R.C and ShangHai Stock Exchange, P.R.C;Department of Computer Science and Engineering, Fudan University, P.R.C and ShangHai Telecom Company, P.R.C;Software Engineering Institute of East China Normal University, P.R.C and Shanghai Key Laboratory of Trustworthy Computering, P.R.C
Venue:
APWeb/WAIM '09 Proceedings of the Joint International Conferences on Advances in Data and Web Management
Year:
2009

Citing 7
Cited 2

LAPACK Users' guide (third ed.)

LAPACK Users' guide (third ed.)
Online Amnesic Approximation of Streaming Time Series

ICDE '04 Proceedings of the 20th International Conference on Data Engineering
Streaming pattern discovery in multiple time-series

VLDB '05 Proceedings of the 31st international conference on Very large data bases
Approximately Processing Multi-granularity Aggregate Queries over Data Streams

ICDE '06 Proceedings of the 22nd International Conference on Data Engineering
Multi-dimensional regression analysis of time-series data streams

VLDB '02 Proceedings of the 28th international conference on Very Large Data Bases
Mining approximate top-k subspace anomalies in multi-dimensional time-series data

VLDB '07 Proceedings of the 33rd international conference on Very large data bases
The MERL motion detector dataset

Proceedings of the 2007 workshop on Massive datasets

Multiple time series anomaly detection based on compression and correlation analysis: a medical surveillance case study

APWeb'12 Proceedings of the 14th Asia-Pacific international conference on Web Technologies and Applications
Probabilistic distance based abnormal pattern detection in uncertain series data

Knowledge-Based Systems

Quantified Score

Hi-index	0.00

Visualization

Abstract

In this paper, we present a method to trace evolution of trend over multiple data streams and detect the abnormal ones. First of all, a definition of trend for single data stream is provided, the advantage of our definition lies in its low time and space cost. Second, we improve a SVD-based method in order to select a pair of optimal initial parameters, then a novel chessboard named sketch is also illustrated aim at adjusting the parameters dynamically. Then, utilizing the skewness of trend distribution, an anomaly detection strategy is briefly introduced. Finally, we implement experiment on a variety of real data sets to illustrate effectiveness and efficiency of our approach.