Seasonality-aware meta-learning framework for improving rainfall prediction accuracy in data-scarce environments 


Vol. 51,  No. 1, pp. 45-54, Jan.  2026
10.7840/kics.2026.51.1.45


PDF Full-Text
  Abstract

This article presents a seasonality-aware meta-learning framework based on MHA-LSTM for enhancing rainfall prediction in data-scarce environments. Utilizing First-Order Model-Agnostic Meta-Learning (FOMAML), the proposed model establishes a mechanism for rapid adaptation to unseen locations with minimal data samples. K-Shot experiments on datasets from 22 South Korean locations confirm that the model achieves high accuracy with only 10 samples. Additionally, a seasonal weighted loss function was integrated to capture the non-linear variability of summer precipitation, ensuring system robustness. This study demonstrates a practical approach to mitigating the cold-start problem in hydrological forecasting and validates the effectiveness of domain-integrated meta-learning for environmental time-series analysis.

  Statistics
Cumulative Counts from November, 2022
Multiple requests among the same browser session are counted as one view. If you mouse over a chart, the values of data points will be shown.


  Related Articles
  Cite this article

[IEEE Style]

D. Kim and Y. Kwon, "Seasonality-aware meta-learning framework for improving rainfall prediction accuracy in data-scarce environments," The Journal of Korean Institute of Communications and Information Sciences, vol. 51, no. 1, pp. 45-54, 2026. DOI: 10.7840/kics.2026.51.1.45.

[ACM Style]

DaeGwang Kim and Young-Woo Kwon. 2026. Seasonality-aware meta-learning framework for improving rainfall prediction accuracy in data-scarce environments. The Journal of Korean Institute of Communications and Information Sciences, 51, 1, (2026), 45-54. DOI: 10.7840/kics.2026.51.1.45.

[KICS Style]

DaeGwang Kim and Young-Woo Kwon, "Seasonality-aware meta-learning framework for improving rainfall prediction accuracy in data-scarce environments," The Journal of Korean Institute of Communications and Information Sciences, vol. 51, no. 1, pp. 45-54, 1. 2026. (https://doi.org/10.7840/kics.2026.51.1.45)
Vol. 51, No. 1 Index

Search
(IN TITLE, AUTHOR, ABSTRACT,KEYWORDS)

Advanced Search ▼

POPULAR KEYWORDS
(TOP 10 KEYWORDS)

Recent Publications
(LAST 3 YEARS)






pISSN : 1226 - 4717
eISSN : 2287 - 3880