Enhancing Cardamom Price Forecasting: Integration of XG Boost Model for a Robust Ensemble Model
Anoop P. S.
*
School of Management and Business Studies, Mahatma Gandhi University Kottayam, India.
Biju M.K.
School of Management and Business Studies, Mahatma Gandhi University Kottayam, India.
Sujith P. S.
School of Computer Science, Mahatma Gandhi University Kottayam, India.
Keerthy T.R.
Cochin University of Science and Technology, Ernakulam, India.
*Author to whom correspondence should be addressed.
Abstract
This study focuses on predicting future cardamom prices in Kerala using data from the Spices Board of Kerala (2014–2024). We propose an ensemble model that integrates the XG Boost machine learning algorithm to enhance predictive accuracy. Our analysis identified daily average price and date as sufficient predictors for forecasting cardamom prices. The results demonstrate that the hybrid ensemble model, particularly with XG Boost, outperforms traditional forecasting methods. These findings highlight the effectiveness of tailored machine learning approaches for complex agricultural markets and suggest a generally stable price structure for cardamom in Kerala, underlining the importance of strategic planning to support farmers' livelihoods.
Keywords: Cardamom pricing, price forecasting, ensemble model, machine learning, XG boost