Volume 28, Issue 4
A Data-Driven Random Subfeature Ensemble Learning Algorithm for Weather Forecasting

Chen Yu, Haochen Li, Jiangjiang Xia, Hanqiuzi WenPingwen Zhang

Commun. Comput. Phys., 28 (2020), pp. 1305-1320.

Published online: 2020-08

Preview Purchase PDF 16 3790
Export citation
  • Abstract

In this paper, the RSEL (Random Subfeature Ensemble Learning) algorithm is proposed to improve the forecast results of weather forecasting. Based on the classical machine learning algorithms, RSEL algorithm integrates random subfeature selection and ensemble learning combination strategy to enhance the diversity of the features and avoid the influence of a small number of unstable outliers generated randomly. Furthermore, the feature engineering schemes are designed for the weather forecast data to make full use of spatial or temporal context. RSEL algorithm is tested by forecasting the wind speed and direction, and it improves the forecast accuracy of traditional methods and has good robustness.

  • Keywords

Weather forecasting, ensemble learning, machine learning, feature engineering.

  • AMS Subject Headings

62P12, 86A10, 93B15, 97M10

  • Copyright

COPYRIGHT: © Global Science Press

  • Email address
  • BibTex
  • RIS
  • TXT
@Article{CiCP-28-1305, author = {Yu , Chen and Li , Haochen and Xia , Jiangjiang and Wen , Hanqiuzi and Zhang , Pingwen}, title = {A Data-Driven Random Subfeature Ensemble Learning Algorithm for Weather Forecasting}, journal = {Communications in Computational Physics}, year = {2020}, volume = {28}, number = {4}, pages = {1305--1320}, abstract = {

In this paper, the RSEL (Random Subfeature Ensemble Learning) algorithm is proposed to improve the forecast results of weather forecasting. Based on the classical machine learning algorithms, RSEL algorithm integrates random subfeature selection and ensemble learning combination strategy to enhance the diversity of the features and avoid the influence of a small number of unstable outliers generated randomly. Furthermore, the feature engineering schemes are designed for the weather forecast data to make full use of spatial or temporal context. RSEL algorithm is tested by forecasting the wind speed and direction, and it improves the forecast accuracy of traditional methods and has good robustness.

}, issn = {1991-7120}, doi = {https://doi.org/ 10.4208/cicp.OA-2020-0006}, url = {http://global-sci.org/intro/article_detail/cicp/18099.html} }
TY - JOUR T1 - A Data-Driven Random Subfeature Ensemble Learning Algorithm for Weather Forecasting AU - Yu , Chen AU - Li , Haochen AU - Xia , Jiangjiang AU - Wen , Hanqiuzi AU - Zhang , Pingwen JO - Communications in Computational Physics VL - 4 SP - 1305 EP - 1320 PY - 2020 DA - 2020/08 SN - 28 DO - http://doi.org/ 10.4208/cicp.OA-2020-0006 UR - https://global-sci.org/intro/article_detail/cicp/18099.html KW - Weather forecasting, ensemble learning, machine learning, feature engineering. AB -

In this paper, the RSEL (Random Subfeature Ensemble Learning) algorithm is proposed to improve the forecast results of weather forecasting. Based on the classical machine learning algorithms, RSEL algorithm integrates random subfeature selection and ensemble learning combination strategy to enhance the diversity of the features and avoid the influence of a small number of unstable outliers generated randomly. Furthermore, the feature engineering schemes are designed for the weather forecast data to make full use of spatial or temporal context. RSEL algorithm is tested by forecasting the wind speed and direction, and it improves the forecast accuracy of traditional methods and has good robustness.

Chen Yu, Haochen Li, Jiangjiang Xia, Hanqiuzi Wen & Pingwen Zhang. (2020). A Data-Driven Random Subfeature Ensemble Learning Algorithm for Weather Forecasting. Communications in Computational Physics. 28 (4). 1305-1320. doi: 10.4208/cicp.OA-2020-0006
Copy to clipboard
The citation has been copied to your clipboard