Scholarlink

Modern agriculture faces increasing challenges such as climate change, water scarcity, and increasing global demand due to population growth. To address these challenges, artificial intelligence (AI) and machine learning techniques are used to analyze diverse and high-quality data. Consider this study on crop monitoring and large-scale yield prediction using machine learning techniques. Features were extracted using Convolutional Neural Networks (CNNs). After feature extraction, principal component analysis (PCA) was performed, and three models were trained using AdaBoost, LightGBM as a decision-tree-based model, and Random Forest as a nonlinear model. These models were then optimized using Bayesian optimization (BO) algorithm. As a result, CNN techniques achieved a remarkable success rate of 67% for AdaBoost, 99.7% for LightGBM, and 99.9% for Random Forest. These results demonstrate that CNN techniques are effective at fine-grained features, contributing to increased crop yields across the board, opening the way for data-driven smart agriculture applications.

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

Fine-Grained Feature Extraction and Yield Forecasting Using CNNs and Optimized Ensemble Models

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Mohammed Khalaf Mahmood,
  • Wisam Dawood Abdullah

摘要

Modern agriculture faces increasing challenges such as climate change, water scarcity, and increasing global demand due to population growth. To address these challenges, artificial intelligence (AI) and machine learning techniques are used to analyze diverse and high-quality data. Consider this study on crop monitoring and large-scale yield prediction using machine learning techniques. Features were extracted using Convolutional Neural Networks (CNNs). After feature extraction, principal component analysis (PCA) was performed, and three models were trained using AdaBoost, LightGBM as a decision-tree-based model, and Random Forest as a nonlinear model. These models were then optimized using Bayesian optimization (BO) algorithm. As a result, CNN techniques achieved a remarkable success rate of 67% for AdaBoost, 99.7% for LightGBM, and 99.9% for Random Forest. These results demonstrate that CNN techniques are effective at fine-grained features, contributing to increased crop yields across the board, opening the way for data-driven smart agriculture applications.