Scholarlink

The rising frequency and severity of extreme weather events are increasingly affecting crop production, underscoring the need for effective risk management strategies. In this context, weather index insurance, leveraging weather data, has emerged as a vital tool in mitigating agricultural risks tied to weather calamities. However, conventional weather-indexed insurance models, based on basic linear relationships, have been hampered by significant basis risks. This study, examining agricultural regions in Illinois and Iowa in the United States, as well as Hunan Province in China, utilizes long-term data on weather, losses, and the insurance market for model construction. The selection of model optimizations was guided by improvements in both utility and Certainty Equivalent Wealth (CEW), employing case-based reasoning to craft a neural network-embedded weather index insurance within an expected utility framework to optimize benefits. The advanced insurance model showcases a 10.23% utility gain and a 3.21% rise in CEW in the Iowa case study. In Hunan, the enhancement in utility is noted at 1.07%, with a CEW increase of 109.98¥/ha. This research delivers an effective and adaptable solution for weather index insurance, poised to significantly improve risk management in agriculture.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Neural Network-Augmented Case-Based Reasoning Framework for Weather Risk Modeling Using Remote Sensing Data

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Yanbing Bai,
  • Laixin Shu,
  • Yunya Wang,
  • Zhengyan Xiao

摘要

The rising frequency and severity of extreme weather events are increasingly affecting crop production, underscoring the need for effective risk management strategies. In this context, weather index insurance, leveraging weather data, has emerged as a vital tool in mitigating agricultural risks tied to weather calamities. However, conventional weather-indexed insurance models, based on basic linear relationships, have been hampered by significant basis risks. This study, examining agricultural regions in Illinois and Iowa in the United States, as well as Hunan Province in China, utilizes long-term data on weather, losses, and the insurance market for model construction. The selection of model optimizations was guided by improvements in both utility and Certainty Equivalent Wealth (CEW), employing case-based reasoning to craft a neural network-embedded weather index insurance within an expected utility framework to optimize benefits. The advanced insurance model showcases a 10.23% utility gain and a 3.21% rise in CEW in the Iowa case study. In Hunan, the enhancement in utility is noted at 1.07%, with a CEW increase of 109.98¥/ha. This research delivers an effective and adaptable solution for weather index insurance, poised to significantly improve risk management in agriculture.