NEURAL NETWORKS FOR DATA QUALITY MONITORING OF TIME SERIES

Augusto Cesar Heluy Dantas, José Manoel de Seixas

2007

Abstract

Time series play an important role in most of large data bases. Much of the information comes in temporal patterns which is often used for decision taking. Problems with missing and noisy data arise when data quality is not monitored, generating losses in many fields such as economy, customer relationship and health management. In this paper we present a neural network based system used to provide data quality monitoring for time series data. The goal of this system is to continuously adapt a neural model for each monitored series, generating a corridor of acceptance for new observations. Each rejected observation may be substituted by its estimated value, so that data quality is improved. A group of four diverse time series was tested and the system proved to be able to detect the induced outliers.

References

  1. Chatfield, C. (1984). Analisys of Time Series. Chapman and Hall.
  2. Dantas, A. C. H. and Seixas, J. M. (2005). Adaptive neural system for financial time series tracking. In Ribeiro, B., editor, ICANNGA - International Conference on Adaptive and Natural Computing Algorithms, Springer Computer Series: Adaptive and Natural Computing Algorithms, pages 421-424. Elsevier.
  3. Dickey, D. A. and Fuller, W. A. (1979). Distributions of the estimators for autoregressive time series with a unit root. Journal of the American Statistical Association, 75:427-431.
  4. Donoho, D. L. (2000). High-dimensional data analysis: The curses and blessings of dimensionality. Lecture for the American Math. Society “Math Challenges of the 21st Century”.
  5. Eckerson, W. W. (2001). Data quality and the bottom line. Technical report, The Data Warehousing Institute.
  6. Economagic (2006). www.economagic.com.
  7. Haykin, S. (1999). Neural Networks - a Comprehensive Foundation. Prentice-Hall, 2nd. edition.
  8. Kaastra, I. and Boyd, M. (1996). Designing a neural network for forecasting financial and economic time series. In Neurocomputing, number 10, pages 215-236. Elsevier.
  9. Montgomery, D. C., Lynwood, A. J., and Gardner, J. S. (1990). Forecasting and Time Series Analysis. McGraw-Hill.
  10. Olsen, J. E. (2003). Data Quality: the Accuracy Dimension. Morgan Kaufmann Publishers.
  11. Phillips, P. C. B. (1987). Time series regression with a unit root. Econometrica, 55(2):277-301.
  12. Stockwiz (2006). Stockwiz web site. www.stockwiz.com.
  13. Yahoo!Finance (2006). http://finance.yahoo.com.
Download


Paper Citation


in Harvard Style

Cesar Heluy Dantas A. and Manoel de Seixas J. (2007). NEURAL NETWORKS FOR DATA QUALITY MONITORING OF TIME SERIES . In Proceedings of the Ninth International Conference on Enterprise Information Systems - Volume 2: ICEIS, ISBN 978-972-8865-89-4, pages 411-415. DOI: 10.5220/0002371004110415


in Bibtex Style

@conference{iceis07,
author={Augusto Cesar Heluy Dantas and José Manoel de Seixas},
title={NEURAL NETWORKS FOR DATA QUALITY MONITORING OF TIME SERIES},
booktitle={Proceedings of the Ninth International Conference on Enterprise Information Systems - Volume 2: ICEIS,},
year={2007},
pages={411-415},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0002371004110415},
isbn={978-972-8865-89-4},
}


in EndNote Style

TY - CONF
JO - Proceedings of the Ninth International Conference on Enterprise Information Systems - Volume 2: ICEIS,
TI - NEURAL NETWORKS FOR DATA QUALITY MONITORING OF TIME SERIES
SN - 978-972-8865-89-4
AU - Cesar Heluy Dantas A.
AU - Manoel de Seixas J.
PY - 2007
SP - 411
EP - 415
DO - 10.5220/0002371004110415