Forecasting Groundnut (Arachis hypogaea L.) Prices in Kurnool Market: A Synergistic Approach Using Machine Learning and Wavelet Analysis

Abstract

Groundnut (Arachis hypogaea L.) is a crop of considerable economic and nutritional significance globally, providing essential nutrients while serving as a vital source of income for millions of farmers. Accurate forecasting of groundnut prices is crucial for ensuring market stability and enabling stakeholders in the agricultural sector to make informed decisions regarding production and marketing. This research focuses on forecasting groundnut prices, specifically in the Kurnool market of Andhra Pradesh. Various advanced forecasting models were employed, including Wavelet-ARIMA, Wavelet-GARCH, Artificial Neural Networks (ANN), and a hybrid ARIMA+ANN approach. The analysis was based on secondary price data collected from AGMARKNET, covering the period from 2010 to 2023, i.e 14 years of data (monthly data). To evaluate the forecasting models, performance metrics such as Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and Mean Absolute Percentage Error (MAPE) were utilised. These metrics are essential for assessing the accuracy and reliability of the models in predicting market trends. Among the models tested, the Wavelet-ARIMA model proved to be the most effective, exhibiting the lowest error metrics and demonstrating a robust ability to capture the underlying price dynamics. The findings highlight the potential of the Wavelet-ARIMA model as a reliable forecasting tool, offering valuable insights for farmers, traders, and policymakers. This research ultimately aims to enhance decision-making and strategic planning within the agricultural sector, contributing to better market outcomes for groundnut.