Scholarlink

<p>Indian Agriculture system is one of the main sources of livelihood, it is necessary to have accurate crop yield Prediction (CYP) to plan agriculture and food security. Nevertheless, existing algorithms show challenges to learn the dependency among the spatial differences in regions and intermittent temporal variations in climatic factors like rainfall and temperature. This paper suggests a new Graph-based Multi-Scale Temporal (GMST) model for CYP. This paper addresses the issue of existing algorithms by learning spatial relations between districts while integrating with the multi-scale learning of the temporal relation. This considers the short-, medium- and long-term yield patterns at the same time. The suggested model is tested on a dataset of environmental, soil, rainfall, and yield data of 32 districts of Rajasthan (2007-2019). The findings indicate that the results have good predictive power with an RMSE, <InlineEquation ID="IEq1"><EquationSource Format="TEX">\(R^2\)</EquationSource><EquationSource Format="MATHML"><math><msup><mi>R</mi><mn>2</mn></msup></math></EquationSource></InlineEquation> and correlation coefficient of 0.090, 0.810, and 0.920, respectively. These results demonstrate that the integration of graph-spatial learning with multi-scale time-varying modelling is very effective in enhancing prediction accuracy, which provides a strong and scalable system of agricultural forecasting and decision-making.</p>

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

Predicting regional agricultural yield in Rajasthan: graph-based multi-scale temporal integrated approach

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月7日

查看全文
  • Majid Hamid Ali,
  • Mamta Kumari,
  • Veena Mittal,
  • Mayank Kumar Jain,
  • Bhawani Singh Rathore

摘要

Indian Agriculture system is one of the main sources of livelihood, it is necessary to have accurate crop yield Prediction (CYP) to plan agriculture and food security. Nevertheless, existing algorithms show challenges to learn the dependency among the spatial differences in regions and intermittent temporal variations in climatic factors like rainfall and temperature. This paper suggests a new Graph-based Multi-Scale Temporal (GMST) model for CYP. This paper addresses the issue of existing algorithms by learning spatial relations between districts while integrating with the multi-scale learning of the temporal relation. This considers the short-, medium- and long-term yield patterns at the same time. The suggested model is tested on a dataset of environmental, soil, rainfall, and yield data of 32 districts of Rajasthan (2007-2019). The findings indicate that the results have good predictive power with an RMSE, \(R^2\)R2 and correlation coefficient of 0.090, 0.810, and 0.920, respectively. These results demonstrate that the integration of graph-spatial learning with multi-scale time-varying modelling is very effective in enhancing prediction accuracy, which provides a strong and scalable system of agricultural forecasting and decision-making.