*Result*: Predicting Cocoa Prices in Ghana Using Machine Learning and Decomposition Based Hybrid Models.

*The contribution of cocoa to the economies of countries, especially producing countries cannot be understated. However, volatility in the prices of commodities affect all stakeholders in the production value chain as it affects planning, policy implementation and mitigation of risks. Thus, it is very crucial to be able to forecast cocoa prices with a high degree of certainty. Hence, this study modeled cocoa prices in Ghana using 15 machine learning models and their corresponding decomposition based hybrid models. The machine learning models incorporated input variables, including interest rates, inflation rates and crude oil prices. Variational mode decomposition (VMD) and ensemble empirical mode decomposition (EEMD) techniques were applied to the data and the 15 models were further used to model the data, thus obtaining 30 hybrid models. The results of the study revealed that among the 15 machine learning models, quantile random forest model was the best for the data set. Generally, EEMD hybrid models performed better than the VMD based models, with EEMD-generalized additive model with splines being the best hybrid model. The findings show that interest rates play a major role in the prediction of cocoa prices in Ghana. This was closely followed by crude oil prices. Hence, it is recommended that policies that would reflect favorable interest rates and crude oil prices are implemented by policy makers. [ABSTRACT FROM AUTHOR]

Copyright of Statistics, Politics & Policy is the property of De Gruyter and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)*