Semi-supervised monitoring of electric load time series for unusual patterns

In this paper we propose a semi-supervised neural network algorithm to identify unusual load patterns in hourly electricity demand time series. In spite of several modeling and forecasting methodologies that have been proposed, there have been limited advancements in monitoring and automatically identifying outlying patterns in such series. This becomes more important considering the difficulty and the cost associated with manual exploration of such data, due to the vast number of observations. The proposed network learns from both labeled and unlabeled patterns, adapting automatically as more data become available. This drastically limits the cost and effort associated with exploring and labeling such data. We compare the proposed method with conventional supervised and unsupervised approaches, demonstrating higher accuracy, robustness and efficacy on empirical electricity load data.

[1]  Guoqiang Peter Zhang,et al.  Neural networks for classification: a survey , 2000, IEEE Trans. Syst. Man Cybern. Part C.

[2]  Gwilym M. Jenkins,et al.  Time series analysis, forecasting and control , 1972 .

[3]  G. Janacek.  Non‐linear Time Series Models in Empirical Finance , 2003 .

[4]  Rob J Hyndman,et al.  Rainbow Plots, Bagplots, and Boxplots for Functional Data , 2010 .

[5]  Chris Chatfield,et al.  The Analysis of Time Series , 1990 .

[6]  Martin T. Hagan,et al.  Neural network design , 1995 .

[7]  B. Silverman,et al.  Functional Data Analysis , 1997 .

[8]  Philip Hans Franses,et al.  Non-Linear Time Series Models in Empirical Finance , 2000 .

[9]  Martial Hebert,et al.  Semi-Supervised Self-Training of Object Detection Models , 2005, 2005 Seventh IEEE Workshops on Applications of Computer Vision (WACV/MOTION'05) - Volume 1.

[10]  David Yarowsky,et al.  Unsupervised Word Sense Disambiguation Rivaling Supervised Methods , 1995, ACL.

[11]  Juan R. Trapero,et al.  Frequency domain methods applied to forecasting electricity markets , 2009 .

[12]  M. Medeiros,et al.  Modeling and forecasting short-term electricity load: A comparison of methods with an application to Brazilian data , 2008 .

[13]  David J. Hand,et al.  A Simple Generalisation of the Area Under the ROC Curve for Multiple Class Classification Problems , 2001, Machine Learning.

[14]  Chris Chatfield,et al.  The Analysis of Time Series: An Introduction , 1981 .

[15]  Rob J Hyndman,et al.  Another look at measures of forecast accuracy , 2006 .

[16]  Leonard J. Tashman,et al.  Out-of-sample tests of forecasting accuracy: an analysis and review , 2000 .

[17]  M. Genton,et al.  Functional Boxplots , 2011 .

[18]  Clive W. J. Granger,et al.  Extracting information from mega‐panels and high‐frequency data , 2008 .

[19]  Bernhard Schölkopf,et al.  Introduction to Semi-Supervised Learning , 2006, Semi-Supervised Learning.

[20]  Robert F. Engle,et al.  The Econometrics of Ultra-High Frequency Data , 1996 .

[21]  Victoria J. Hodge,et al.  A Survey of Outlier Detection Methodologies , 2004, Artificial Intelligence Review.

[22]  P. McSharry,et al.  A comparison of univariate methods for forecasting electricity demand up to a day ahead , 2006 .

[23]  Sridhar Ramaswamy,et al.  Efficient algorithms for mining outliers from large data sets , 2000, SIGMOD '00.

[24]  Silja Meyer-Nieberg,et al.  Electric load forecasting methods: Tools for decision making , 2009, Eur. J. Oper. Res..

[25]  Philip K. Chan,et al.  Modeling multiple time series for anomaly detection , 2005, Fifth IEEE International Conference on Data Mining (ICDM'05).

[26]  P. A. Blight.  The Analysis of Time Series: An Introduction , 1991 .

[27]  Gwilym M. Jenkins,et al.  Time series analysis, forecasting and control , 1971 .

[28]  Nikolaos Kourentzes,et al.  Feature selection for time series prediction - A combined filter and wrapper approach for neural networks , 2010, Neurocomputing.