Estimated electricity consumption for new customers based on neighborhood consumption

• Carolina L. S. Cipriano UFMA
• Weldson A. Corrêa UFMA
• Arthur G. S. Fernandes UFMA
• Mayara G. Silva UFMA
• Stelmo Netto UFMA
• Eliana Monteiro UFMA
• Marcia Izabel A. da Silva UFMA
• Aristófanes C. Silva UFMA

Abstract

Electricity consumption forecasting is currently an area of major interest for most power companies. Despite being a trend in this area, forecasting can be very challenging and even impractical, especially for consumers with little or nonexistent consumption history. We propose in this work an alternative electricity consumption prediction model for consumers without consumption history. The proposed model is based on the x-means algorithm, which uses the k-nearest neighbours’ consumptions to determine consumer groups, and stochastic gradient descent regressor to create a consumption estimate. The proposed method achieved promising results, in which we highlight the mean absolute percentage error of 38.76% and Theil Inequality Coefficient of 29.56%.

Keywords: Prediction, Power Consumption, nearest k-neighbors

References

Kim, M., Park, S., Lee, J., Joo, Y., and Choi, J. K. (2017). Learning-based adaptive imputation methodwith knn algorithm for missing power data. Energies, 10(10).

Leuthold, R. M. (1975). On the use of theil’s inequality coefficients. American Journal of Agricultural Economics, 57(2):344–346.

Lora, A. T., Santos, J. R., Santos, J. R., Ramos, J. L. M., and Exposito, A. G. (2002). Electricity market price forecasting: Neural networks versus weighted-distance k nearest neighbours. In Hameurlain, A., Cicchetti, R., and Traunm¨uller, R., editors, Database and Expert Systems Applications, pages 321–330, Berlin, Heidelberg. Springer Berlin Heidelberg.

McLoughlin, F., Duffy, A., and Conlon, M. (2015). A clustering approach to domestic electricity load profile characterisation using smart metering data. Applied energy, 141:190–199.

Pelleg, D., Moore, A. W., et al. (2000). X-means: extending k-means with efficient estimation of the number of clusters. In Icml, volume 1, pages 727–734.

Poloczek, J., Treiber, N. A., and Kramer, O. (2014). Knn regression as geo-imputation method for spatio-temporal wind data. In de la Puerta, J. G., Ferreira, I. G., Bringas,

P. G., Klett, F., Abraham, A., de Carvalho, A. C., Herrero, A´ ., Baruque, B., Quinti´an, H., and Corchado, E., editors, International Joint Conference SOCO’14- CISIS’14-ICEUTE’14, pages 185–193, Cham. Springer International Publishing.

Ruder, S. (2016). An overview of gradient descent optimization algorithms. arXiv preprint arXiv:1609.04747.

Wahid, F. and Kim, D. (2016). A prediction approach for demand analysis of energy consumption using k-nearest neighbor in residential buildings. International Journal of Smart Home, 10:97–108.

Williamson, D. F., Parker, R. A., and Kendrick, J. S. (1989). The box plot: a simple visual method to interpret data. Annals of internal medicine, 110(11):916–921.

Yorucu, V. (2003). The analysis of forecasting performance by using time series data for two mediterranean islands. Review of Social, Economic & Business Studies, 2:175– 196.