Scholarlink

<p>Pomegranate is a&#xa0;nutritionally rich and economically vital fruit crop. It is recognized as one of India’s vital cash crops with Maharashtra leading in its cultivation. Despite its significance, research on price forecasting for pomegranate remains limited, highlighting the need for accurate and reliable forecasting methods. This study explores a&#xa0;diverse range of stochastic, machine learning, and deep learning models to forecast pomegranate prices in major South Indian markets. Various evaluation metrics such as root mean square error (RMSE), mean absolute percentage error (MAPE), mean absolute error (MAE), root mean square percentage error (RMSPE), and median absolute percentage error (median APE) are used to evaluate the model performance. The results demonstrate the superior performance of deep learning models, including convolutional neural network (CNN), recurrent neural network (RNN), long short-term memory (LSTM), and gated recurrent unit (GRU), compared to traditional and machine learning models. In Bangalore, the CNN model achieved the lowest MAPE of 10.27%, indicating the best performance. For Chennai, the GRU model outperformed others with a&#xa0;MAPE of 11.81%. In Hyderabad, the RNN model delivered the best result with a&#xa0;MAPE of 19.19%, while in Trivandrum, the LSTM model recorded the lowest MAPE of 14.6%. This highlights the effectiveness of deep learning approaches in price forecasting, further emphasizing that model performance depends on the underlying characteristics of the data. The study provides valuable insights for farmers, stakeholders, and policymakers to better manage the price volatility in pomegranate across different marketing levels.</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 of Deep Learning Approaches for Price Forecasting: A Case Study on Pomegranate

  • 期刊论文

  • 发表日期：

    2025年6月23日

查看全文
  • S. Vishnu Shankar,
  • V. Lavanya,
  • P. Pandiyaraj,
  • R. Ajaykumar,
  • Adit Kumar Yadav

摘要

Pomegranate is a nutritionally rich and economically vital fruit crop. It is recognized as one of India’s vital cash crops with Maharashtra leading in its cultivation. Despite its significance, research on price forecasting for pomegranate remains limited, highlighting the need for accurate and reliable forecasting methods. This study explores a diverse range of stochastic, machine learning, and deep learning models to forecast pomegranate prices in major South Indian markets. Various evaluation metrics such as root mean square error (RMSE), mean absolute percentage error (MAPE), mean absolute error (MAE), root mean square percentage error (RMSPE), and median absolute percentage error (median APE) are used to evaluate the model performance. The results demonstrate the superior performance of deep learning models, including convolutional neural network (CNN), recurrent neural network (RNN), long short-term memory (LSTM), and gated recurrent unit (GRU), compared to traditional and machine learning models. In Bangalore, the CNN model achieved the lowest MAPE of 10.27%, indicating the best performance. For Chennai, the GRU model outperformed others with a MAPE of 11.81%. In Hyderabad, the RNN model delivered the best result with a MAPE of 19.19%, while in Trivandrum, the LSTM model recorded the lowest MAPE of 14.6%. This highlights the effectiveness of deep learning approaches in price forecasting, further emphasizing that model performance depends on the underlying characteristics of the data. The study provides valuable insights for farmers, stakeholders, and policymakers to better manage the price volatility in pomegranate across different marketing levels.