Combined General Vector Machine for Single Point Electricity Load Forecast

Authors

Binbin Yong, University of WollongongFollow
Yongqiang Wei
Jun Shen, University of WollongongFollow
Fucun Li, University of WollongongFollow
Xuetao Jiang
Qingguo ZhouFollow

RIS ID

134655

Publication Details

Yong, B., Wei, Y., Shen, J., Li, F., Jiang, X. & Zhou, Q. (2020). Combined General Vector Machine for Single Point Electricity Load Forecast. The 9th International Conference on Frontier Computing (FC2019) Theory, Technologies and Applications (pp. 284-290). LNEE: Springer.

Abstract

General Vector Machine (GVM) is a newly proposed machine learning model, which is applicable to small samples forecast scenarios. In this paper, the GYM is applied into electricity load fore­cast based on single point modeling method. Meanwhile, traditional time series forecast models, including back propagation neural network (BPNN), Support Vector Machine (SVM) and Autoregressive Integrated Moving Average Model ( ARIMA), are also experimented for single point electricity load forecast. Further, the combined model based on GYM, BPNN, SVM and ARIMA are proposed and verified. Results show that GYM performs better than these traditional models, and the combined model outperforms any other single models for single point electricity load forecast.