Scholarlink

<p>Weather prediction remains a challenging task for meteorological researchers, despite advancements in forecasting technologies. This manuscript proposes an Enhanced Smart Weather Prediction through Advanced Atmospheric Analysis and Forecasting Techniques (SWP-AAFT-BSNN) using Binarized Spiking Neural Networks (BSNN). The method begins by collecting data from a rain prediction database, which is pre-processed using the Sigma-Mixed Unscented Kalman Filter (SMUKF) to remove noise. Optimal feature selection is performed using the Wader Hunt Optimization Algorithm (WHOA). The selected features are fed into BSNN for differentiating weather forecasting as rain and no rain. The Leaf in Wind Optimization Algorithm (LWOA) is introduced to optimize the BSNN weight parameters. The proposed SWP-AAFT-BSNN approach is evaluated under the performance metrics, like accuracy, precision, sensitivity, F1-score, MAE, RMSE, and computational time. The results demonstrate a significant improvement, with accuracy increasing by 19.46%, 24.77%, and 29.50%, RMSE reducing by 18.21%, 35.82%, and 28.64%, and computational time decreasing by 16.64%, 28.46%, and 30.44% when compared to the existing methods.</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

Enhanced smart weather prediction through advanced atmospheric analysis and forecasting techniques using binarized spiking neural networks

  • 期刊论文

  • 发表日期：

    2025年1月25日

查看全文
  • M. Amanullah,
  • K. Ananthajothi,
  • D. Divya

摘要

Weather prediction remains a challenging task for meteorological researchers, despite advancements in forecasting technologies. This manuscript proposes an Enhanced Smart Weather Prediction through Advanced Atmospheric Analysis and Forecasting Techniques (SWP-AAFT-BSNN) using Binarized Spiking Neural Networks (BSNN). The method begins by collecting data from a rain prediction database, which is pre-processed using the Sigma-Mixed Unscented Kalman Filter (SMUKF) to remove noise. Optimal feature selection is performed using the Wader Hunt Optimization Algorithm (WHOA). The selected features are fed into BSNN for differentiating weather forecasting as rain and no rain. The Leaf in Wind Optimization Algorithm (LWOA) is introduced to optimize the BSNN weight parameters. The proposed SWP-AAFT-BSNN approach is evaluated under the performance metrics, like accuracy, precision, sensitivity, F1-score, MAE, RMSE, and computational time. The results demonstrate a significant improvement, with accuracy increasing by 19.46%, 24.77%, and 29.50%, RMSE reducing by 18.21%, 35.82%, and 28.64%, and computational time decreasing by 16.64%, 28.46%, and 30.44% when compared to the existing methods.