Scholarlink

<p>Farmers occasionally cultivate crops that result in less yield and lead to the wastage of their land, labor, and time. The situation becomes more challenging in developing countries where the need to feed the large population with limited crop production is increasing daily. Thus, implementing a crop recommendation system can help alleviate this challenge. A crop recommendation system is a modern precision farming practice, which involves gathering, analyzing, and utilizing individual, spatial, and temporal data with the assistance of diverse machine learning and deep learning models. Precision farming represents a contemporary farming approach leveraging soil data, including soil features, types, and crop yield information, to recommend the most suitable crops based on specific parameters of each farm. Currently, due to a lack of appropriate features, the accuracy of the recommendation system is compromised, and processing the climate dataset is time-consuming. This study employs a Long Short Term Memory Network Ensemble via Adaptive Weighting (AWLSTM) method alongside the Dynamic Salp Swarm Algorithm (DSSA) to ensure accurate crop predictions and efficient recommendations. Initially, climate data is gathered, encompassing various variables affecting rainfall and agricultural yield in specific locations. Subsequently, pre-processing is conducted to refine the quality of the input data. To achieve accurate prediction outcomes, the DSSA algorithm is employed to select the most relevant features based on optimal fitness values. Crop predictions are then conducted using the AWLSTM method. Experimental results demonstrate that DSSA-AWLSTM provides better results in terms of precision, recall, and execution time.</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

An Effective Crop Recommendation System Using A Dynamic Salp Swarm Algorithm with Adaptive Weighting Based LSTM Network

  • 期刊论文

  • 发表日期：

    2025年6月16日

查看全文
  • Parijata Majumdar,
  • Sanjoy Mitra,
  • Debashis De

摘要

Farmers occasionally cultivate crops that result in less yield and lead to the wastage of their land, labor, and time. The situation becomes more challenging in developing countries where the need to feed the large population with limited crop production is increasing daily. Thus, implementing a crop recommendation system can help alleviate this challenge. A crop recommendation system is a modern precision farming practice, which involves gathering, analyzing, and utilizing individual, spatial, and temporal data with the assistance of diverse machine learning and deep learning models. Precision farming represents a contemporary farming approach leveraging soil data, including soil features, types, and crop yield information, to recommend the most suitable crops based on specific parameters of each farm. Currently, due to a lack of appropriate features, the accuracy of the recommendation system is compromised, and processing the climate dataset is time-consuming. This study employs a Long Short Term Memory Network Ensemble via Adaptive Weighting (AWLSTM) method alongside the Dynamic Salp Swarm Algorithm (DSSA) to ensure accurate crop predictions and efficient recommendations. Initially, climate data is gathered, encompassing various variables affecting rainfall and agricultural yield in specific locations. Subsequently, pre-processing is conducted to refine the quality of the input data. To achieve accurate prediction outcomes, the DSSA algorithm is employed to select the most relevant features based on optimal fitness values. Crop predictions are then conducted using the AWLSTM method. Experimental results demonstrate that DSSA-AWLSTM provides better results in terms of precision, recall, and execution time.