Sign in Subscribe
Agricultural industry

Deep Learning Enhances Predictive Accuracy in Agricultural Prices

Strategic Insights

— 1 min read

New research from the Indian Council of Agricultural Research (ICAR) Indian Agricultural Statistics Research Institute has demonstrated the effectiveness of deep learning methodologies in handling the complex patterns of agricultural price datasets. The study compared the performance of various deep learning models, including Convolutional Neural Network (CNN), Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM), and Gated Recurrent Unit (GRU), against feed-forward Artificial Neural Networks (ANN) in predicting potato prices across four markets in India. The results showed that the GRU model performed best in two markets, while LSTM and CNN models showed superior performance in the other two markets.

Key Takeaways:

  • The study demonstrated the effectiveness of deep learning methodologies in handling the complex patterns of agricultural price datasets.
  • The GRU model performed best in two markets, achieving a Mean Absolute Percentage Error (MAPE) of 16.26% and 6.09% in Chandigarh and Delhi, respectively.
  • LSTM model showed superior performance in the Dehradun market with a MAPE of 17.81%, while CNN model performed best in Shimla market with a MAPE of 12.53%.
  • The error percentage of deep learning models were remarkably low when compared to the machine learning model.
  • The study suggests that deep learning can enhance predictive accuracy and understanding of the complexities of agricultural prices.
  • The research highlights the potential of neural networks in solving the intricate nature of agricultural price data.
  • The study was conducted using monthly potato price data collected from the National Horticultural Board across four distinct markets in India.

Statistics:

  • The GRU model performed best in two markets, achieving a MAPE of 16.26% and 6.09% in Chandigarh and Delhi, respectively.
  • LSTM model showed superior performance in the Dehradun market with a MAPE of 17.81%.
  • CNN model performed best in Shimla market with a MAPE of 12.53%.
  • The error percentage of deep learning models were remarkably low, with a MAPE of 13.44% on average.

Sources:

  • "Comparative study of neural network variants for potato (Solanum tuberosum) price modeling." The Indian Journal of Agricultural Sciences, 2025, 95(8).
  • Indian Council of Agricultural Research (ICAR) Indian Agricultural Statistics Research Institute.
  • S Vishnu Shankar, Indian Council of Agricultural Research (ICAR) Indian Agricultural Statistics Research Institute, New Delhi.
  • Ranjit Kumar Paul, Md Yeasin, Patil Santosh Ganapati, Indian Council of Agricultural Research (ICAR) Indian Agricultural Statistics Research Institute.
  • Journal of Engineering.

Read more