Table of Content

Open Access iconOpen Access

ARTICLE

crossmark

Analysis and Application of the Spatio-Temporal Feature in Wind Power Prediction

Ruiguo Yu1,2, Zhiqiang Liu1,2, Jianrong Wang1,3, Mankun Zhao1,2, Jie Gao1,3, Mei Yu1,3,*

1 School of Computer Science and Technology, Tianjin University, Tianjin 300350, China
{rgyu; tjubeisong; wjr; zmk; gaojie; yumei}@tju.edu.cn
2 Tianjin Key Laboratory of Cognitive Computing and Application, Tianjin 300350, China
3 Tianjin Key Laboratory of Advanced Networking, Tianjin 300350, China

* Corresponding Author: E-mail: email

Computer Systems Science and Engineering 2018, 33(4), 267-274. https://doi.org/10.32604/csse.2018.33.267

Abstract

The spatio-temporal feature with historical wind power information and spatial information can effectively improve the accuracy of wind power prediction, but the role of the spatio-temporal feature has not yet been fully discovered. This paper investigates the variance of the spatio-temporal feature. Based on this, a hybrid machine learning method for wind power prediction is designed. First, the training set is divided into several groups according to the variance of the input pattern, and then each group is used to train one or more predictors respectively. Multiple machine learning methods, such as the support vector machine regression and the decision tree, are used in the proposed method. Second, all the trained predictors are adopted to make predictions for a sample, and the results generated from these predictors will be combined by an optimized combination method based on the variance. The experimental results based on the NREL dataset show that the method adopted in this paper can achieve a better performance than the stage-of-the-art approaches.

Keywords


Cite This Article

R. Yu, Z. Liu, J. Wang, M. Zhao, J. Gao et al., "Analysis and application of the spatio-temporal feature in wind power prediction," Computer Systems Science and Engineering, vol. 33, no.4, pp. 267–274, 2018.



cc This work is licensed under a Creative Commons Attribution 4.0 International License , which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
  • 1165

    View

  • 946

    Download

  • 0

    Like

Share Link