A Fuzzy Genetic Algorithm Classifier: The Impact of Time-Series Load Data Temporal Dimension on Classification Performance

Utilization of machine learning algorithms in time-series data analysis is crucial to effective decision making in today's dynamic and competitive environment. One data type of growing interest is the electricity consumer load profile (LP) data. Owing to advances in the smart grid, immense amount of LP data became available to policymakers as potential to improving the electricity sector. Due to the growing size and volatile nature of LP data, development and evaluation of clustering approaches has been of high demand in recent energy research, whereas the classification techniques receive less attention. This study is the first to address the effect of LP time-series data temporal dimension on the classification performance using the most popular classification algorithms in machine learning including decision trees, support vector machines (SVM), discriminant analysis, and ensemble methods. Results indicate a decline in the classification accuracy as the temporal dimension increases. Accordingly, this study proposes a fuzzy classification heuristic-based method inspired by the genetic algorithm (GA) which proves to maintain robustness against high temporal dimensions. The results are assessed using real data from industrial consumers with 420 daily LPs and 93 weekly LPs.

[1]  J. Ross Quinlan,et al.  Induction of Decision Trees , 1986, Machine Learning.

[2]  Anil K. Jain.  Data clustering: 50 years beyond K-means , 2008, Pattern Recognit. Lett..

[3]  Helge V. Larsen,et al.  Long-term forecasting of hourly electricity load: Identification of consumption profiles and segmentation of customers , 2013 .

[4]  Mohammad Kazem Sheikh-El-Eslami,et al.  A three-stage strategy for optimal price offering by a retailer based on clustering techniques , 2010 .

[5]  K. Steemers,et al.  A method of formulating energy load profile for domestic buildings in the UK , 2005 .

[6]  Zita Vale,et al.  A data-mining-based methodology to support MV electricity customers’ characterization , 2015 .

[7]  James J. Chen,et al.  Classification by ensembles from random partitions of high-dimensional data , 2007, Comput. Stat. Data Anal..

[8]  Lingfeng Wang,et al.  Electricity theft: Overview, issues, prevention and a smart meter based approach to control theft , 2011 .

[9]  Michalis Vazirgiannis,et al.  Clustering algorithms and validity measures , 2001, Proceedings Thirteenth International Conference on Scientific and Statistical Database Management. SSDBM 2001.

[10]  Antonio Cezar de Castro Lima,et al.  Typification of load curves for DSM in Brazil for a smart grid environment , 2015 .

[11]  Gianfranco Chicco,et al.  Overview and performance assessment of the clustering methods for electrical load pattern grouping , 2012 .

[12]  Ioannis Panapakidis,et al.  Evaluation of the performance of clustering algorithms for a high voltage industrial consumer , 2015, Eng. Appl. Artif. Intell..

[13]  Desire L. Massart,et al.  Comparison of regularized discriminant analysis linear discriminant analysis and quadratic discriminant analysis applied to NIR data , 1996 .

[14]  Jean-Michel Poggi,et al.  OPTIMIZED CLUSTERS FOR DISAGGREGATED ELECTRICITY LOAD FORECASTING , 2010 .

[15]  Fábio Gonçalves Jota,et al.  Building load management using cluster and statistical analyses , 2011 .

[16]  Hongseok Kim,et al.  A framework for baseline load estimation in demand response: Data mining approach , 2014, 2014 IEEE International Conference on Smart Grid Communications (SmartGridComm).

[17]  C. S. Chen,et al.  Application of load survey systems to proper tariff design , 1997 .

[18]  Vitaly Ford,et al.  Clustering of smart meter data for disaggregation , 2013, 2013 IEEE Global Conference on Signal and Information Processing.

[19]  Aidan Duffy,et al.  Evaluation of time series techniques to characterise domestic electricity demand , 2013 .

[20]  Ahmed Abdulaal,et al.  Electric load pattern classification for demand-side management planning: A hybrid approach , 2015 .

[21]  A. Atiya,et al.  Learning with Kernels: Support Vector Machines, Regularization, Optimization, and Beyond , 2005, IEEE Transactions on Neural Networks.

[22]  Mohammad Kazem Sheikh-El-Eslami,et al.  An annual framework for clustering-based pricing for an electricity retailer , 2010 .

[23]  J. Razmi,et al.  Forecasting electricity consumption by clustering data in order to decline the periodic variable’s affects and simplification the pattern , 2009 .

[24]  P. Rousseeuw.  Silhouettes: a graphical aid to the interpretation and validation of cluster analysis , 1987 .

[25]  Ahmed Abdulaal,et al.  A linear optimization based controller method for real-time load shifting in industrial and commercial buildings , 2016 .

[26]  Ahmed Abdulaal,et al.  Electric load pattern classification using parameter estimation, clustering and artificial neural networks , 2015 .

[27]  D.J. King,et al.  Electricity load profile classification using Fuzzy C-Means method , 2008, 2008 43rd International Universities Power Engineering Conference.