Scholarlink

<p>Potato is a vital staple commodity with substantial economic significance across various agricultural sectors. However, its market prices are highly volatile, leading to challenges in accurate forecasting due to frequent and unpredictable fluctuations. Such volatility adversely impacts farmers’ incomes and consumer affordability, thereby complicating agricultural planning and policy formulation. Hence, accurate and reliable price forecasting is essential to mitigate risks and stabilize agricultural market stability. This study evaluates the effectiveness of wavelet-based hybrid models in forecasting monthly wholesale potato prices from January 2006 to December 2022 in major South Indian markets such as Chennai, Trivandrum, Bangalore, and Hyderabad. Wavelet decomposition was employed to capture both long-term trends and short-term fluctuations from the price series, thereby enhancing the predictive accuracy of the models. Model performance is assessed using evaluation metrics such as root mean square error (RMSE), mean absolute percentage error (MAPE), and mean absolute error (MAE), and further validated through the Diebold-Mariano (DM) test. The results reveal that the wavelet-autoregressive integrated moving average model achieved the highest forecasting accuracy for Chennai and Trivandrum markets, with MAPE values of 10.09% and 10.46%, respectively. The wavelet-artificial neural networks model demonstrated superior performance for Bangalore and Hyderabad markets, with MAPE values of 22.02% and 7.93%, respectively. The DM test confirms the superior performance of the wavelet-based models over the benchmark models across different markets. These findings highlight the efficiency of wavelet-enhanced models in capturing the complex dynamics of agricultural price series. The study offers valuable insights for stakeholders including farmers, traders and policymakers by facilitating data-driven decisions to manage price risks. Furthermore, the study highlights the critical need for implementing robust and adaptive pricing policies to mitigate market volatility, safeguard farmer livelihoods and ensure fair and stable returns. Specifically, the adoption of a fixed price policy for potatoes is recommended as a proactive governmental measure to enhance economic resilience in this highly volatile sector. Future research should explore more advanced hybrid models and real-time forecasting frameworks to further improve adaptability and predictive accuracy in volatile agricultural markets.</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

Comparative Analysis Using a Wavelet–enhanced Time Series Approach for Potato Price Modelling

  • 期刊论文

  • 发表日期：

    2025年10月29日

查看全文
  • V Lavanya,
  • Gangai Selvi R,
  • Patil Santosh Ganapati,
  • J Auxcilia,
  • V Karthick,
  • S Vishnu Shankar

摘要

Potato is a vital staple commodity with substantial economic significance across various agricultural sectors. However, its market prices are highly volatile, leading to challenges in accurate forecasting due to frequent and unpredictable fluctuations. Such volatility adversely impacts farmers’ incomes and consumer affordability, thereby complicating agricultural planning and policy formulation. Hence, accurate and reliable price forecasting is essential to mitigate risks and stabilize agricultural market stability. This study evaluates the effectiveness of wavelet-based hybrid models in forecasting monthly wholesale potato prices from January 2006 to December 2022 in major South Indian markets such as Chennai, Trivandrum, Bangalore, and Hyderabad. Wavelet decomposition was employed to capture both long-term trends and short-term fluctuations from the price series, thereby enhancing the predictive accuracy of the models. Model performance is assessed using evaluation metrics such as root mean square error (RMSE), mean absolute percentage error (MAPE), and mean absolute error (MAE), and further validated through the Diebold-Mariano (DM) test. The results reveal that the wavelet-autoregressive integrated moving average model achieved the highest forecasting accuracy for Chennai and Trivandrum markets, with MAPE values of 10.09% and 10.46%, respectively. The wavelet-artificial neural networks model demonstrated superior performance for Bangalore and Hyderabad markets, with MAPE values of 22.02% and 7.93%, respectively. The DM test confirms the superior performance of the wavelet-based models over the benchmark models across different markets. These findings highlight the efficiency of wavelet-enhanced models in capturing the complex dynamics of agricultural price series. The study offers valuable insights for stakeholders including farmers, traders and policymakers by facilitating data-driven decisions to manage price risks. Furthermore, the study highlights the critical need for implementing robust and adaptive pricing policies to mitigate market volatility, safeguard farmer livelihoods and ensure fair and stable returns. Specifically, the adoption of a fixed price policy for potatoes is recommended as a proactive governmental measure to enhance economic resilience in this highly volatile sector. Future research should explore more advanced hybrid models and real-time forecasting frameworks to further improve adaptability and predictive accuracy in volatile agricultural markets.