search for


Predictability model of the sea ice extent based on a machine learning technique
Journal of the Korean Data & Information Science Society 2023;34:331-40
Published online March 31, 2023;
© 2023 Korean Data and Information Science Society.

Young Eun Jeon1 · Suk-Bok Kang2 · Jung-In Seo3

12Department of Statistics, Yeungnam University
3Department of Information Statistics, Andong National University
Correspondence to: This work was supported by the Ministry of Education of the Republic of Korea and the National Research Foundation of Korea (NRF-2021S1A5A8063350).
1 Graduate student, Department of Statistics, Yeungnam University, Gyeongsan 38541, Korea.
2 Professor, Department of Statistics, Yeungnam University, Gyeongsan 38541, Korea.
3 Assistant professor, Department of Information Statistics, Andong National University, Andong 36729, Korea. E-mail :
Received January 2, 2023; Revised January 28, 2023; Accepted February 2, 2023.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License ( which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Recent climate disasters such as heat waves and heavy rains have emerged in the United States. These climate disasters are mainly caused by melting sea ice, an important indicator of maintaining the global average temperature. As sea ice continues to decrease due to global warming, this is the time when a strategy to respond to climate disasters caused by a decrease in the sea ice extent is more necessary than ever. As an aid to this, this study provides predictive modeling using a tree-based machine learning approach to predict the decrease in the Arctic sea ice extent. Specifically, the prediction is accomplished through three procedures: First, to impute the missing values of the Arctic sea ice extent, an imputation method using a Kalman smoothing technique is implemented. In addition, various features including the Fourier terms representing a seasonal variation are extracted and generated. Finally, to resolve the drawback of the tree-based machine learning techniques which cannot capture trends, a hybrid strategy based on the combination of the statistical and tree-based machine learning techniques is used. The provided hybrid strategy is thought to be a very good guideline for applying a tree-based machine learning technique to the Arctic sea ice extent.
Keywords : Arctic sea ice extent, Fourier terms, hybrid strategy, machine learning.