Scholarlink

<p>In recent years, artificial intelligence (AI) technologies, including machine learning (ML) and deep learning (DL) methods such as convolutional neural networks (CNNs) and recurrent neural networks (RNNs), have emerged as a&#xa0;powerful tool for addressing complex problems in agriculture, particularly in yield forecasting. Applications of these techniques are most effective among AI models for crop yield prediction. This study explores the performance of ML and DL techniques in predicting apple yields within the horticulture sector. Traditional methods of yield forecasting rely heavily on historical data, statistical models, and expert knowledge, often limiting the accuracy of predictions. In contrast, AI-based models capture intricate patterns in large datasets, including weather conditions, soil properties, and crop management practices. In this case study, we employ ML and DL models to predict apple yields using morphological data collected from 300 randomly selected ‘Golden Delicious’ apple trees from well-established orchards in the Shimla district of Himachal Pradesh, India. The performance of these models is compared to traditional statistical techniques. Our findings suggest that support vector regression (SVR) offers significant potential for enhancing yield prediction with consistency in both training and test datasets in apple farming, enabling growers to make more informed decisions regarding resource allocation and crop management.</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

A Comparative Study on the Predictive Ability of Machine Learning- and Deep Learning-Based Yield Prediction Model in Horticulture: A Case Study of Apple

  • 期刊论文

  • 发表日期：

    2025年6月23日

查看全文
  • Anju Sharma,
  • Pankaj Das,
  • Samir Barman,
  • Satish K. Sharma,
  • Bulbul Ahmed,
  • Trina Adhikary

摘要

In recent years, artificial intelligence (AI) technologies, including machine learning (ML) and deep learning (DL) methods such as convolutional neural networks (CNNs) and recurrent neural networks (RNNs), have emerged as a powerful tool for addressing complex problems in agriculture, particularly in yield forecasting. Applications of these techniques are most effective among AI models for crop yield prediction. This study explores the performance of ML and DL techniques in predicting apple yields within the horticulture sector. Traditional methods of yield forecasting rely heavily on historical data, statistical models, and expert knowledge, often limiting the accuracy of predictions. In contrast, AI-based models capture intricate patterns in large datasets, including weather conditions, soil properties, and crop management practices. In this case study, we employ ML and DL models to predict apple yields using morphological data collected from 300 randomly selected ‘Golden Delicious’ apple trees from well-established orchards in the Shimla district of Himachal Pradesh, India. The performance of these models is compared to traditional statistical techniques. Our findings suggest that support vector regression (SVR) offers significant potential for enhancing yield prediction with consistency in both training and test datasets in apple farming, enabling growers to make more informed decisions regarding resource allocation and crop management.