An Extreme Gradient Boosting Algorithm for Short-Term Load Forecasting Using Power Grid Big Data
Directed at the problem of more and increasing data types and volume in power grid, a short-term power load forecasting algorithm based on big data and Extreme Gradient Boosting (XGBoost) is proposed, based on the analysis of power grid load big data low. The algorithm includes the following steps. First, the outlier data and missing data are preprocessed. Then, the K-means algorithm is used to cluster the load big data of the power grid. Finally, The XGBoost algorithm was used to train the load forecasting model, based on the impact of historical load, calendar effect and meteorological factors on the load. Simulation results show that compared with support vector machine, random forest and decision tree, the proposed algorithm has a higher prediction accuracy and smoother prediction error, with smaller mean absolute percentage error, mean absolute error and relative error.
KeywordsPower grid big data Power load forecasting Extreme gradient boosting algorithm Cluster analysis
This research was supported by the National Natural Science Foundation of China (Grant No. 91538201); Taishan scholar project special fund project (Grant No. Ts201511020).
Limin Zhang, Qiang Guo design experiments and collect data. Liqiang Ren and Haipeng Wang conducted a case study and analyzed the results. Liqiang Ren wrote this paper. All authors read and approved the final manuscript.
Conflicts of Interest
The authors declare no conflict of interest.
- 4.M.Y. Cho, J.C. Hwang, C.S. Chen, Customer short term load forecasting by using ARIMA transfer function model, in International Conference on Energy Management and Power Delivery, 1995. Proceedings of EMPD (IEEE, 1995), pp. 317–322Google Scholar
- 7.W. Baoyi, Z. Shuo, Z. Shaomin, Distributed power load forecasting algorithm based on cloud computing and extreme learning machines. Power Syst. Technol. 38(2), 526–531 (2014)Google Scholar
- 9.L. Yuhua, C. Hong, G. Kun et al., Research on load forecasting based on random forest algorithm. CEA 52(23), 236–243 (2016)Google Scholar
- 10.T. Chen, C. Guestrin, XGBoost: a scalable tree boosting system, in Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, San Francisco, CA, USA, 13–17 August 2016, pp. 785–794Google Scholar
- 13.M. Liu, The Power System Load Forecasting Research Based on Wavelet and Neural Network Theory (Nanjing University of Science and Technology, Nanjing, 2012)Google Scholar
- 14.R.J. Broderick, J.R. Williams, Clustering methodology for classifying distribution feeders, in Photovoltaic Specialists Conference (IEEE, 2013), pp. 1706–1710Google Scholar