Forecasting commodity prices using long-short-term memory neural networks

This paper applies a recurrent neural network (RNN) method to forecast cotton and oil prices. We show how these new tools from machine learning, particularly Long-Short Term Memory (LSTM) models, complement traditional methods. Our results show that machine learning methods fit reasonably well with...

Descripción completa

Detalles Bibliográficos
Autores principales: Ly, Racine, Traoré, Fousseini, Dia, Khadim
Formato: Artículo preliminar
Lenguaje:Inglés
Publicado: International Food Policy Research Institute 2021
Materias:
Acceso en línea:https://hdl.handle.net/10568/143474
_version_ 1855524609296695296
author Ly, Racine
Traoré, Fousseini
Dia, Khadim
author_browse Dia, Khadim
Ly, Racine
Traoré, Fousseini
author_facet Ly, Racine
Traoré, Fousseini
Dia, Khadim
author_sort Ly, Racine
collection Repository of Agricultural Research Outputs (CGSpace)
description This paper applies a recurrent neural network (RNN) method to forecast cotton and oil prices. We show how these new tools from machine learning, particularly Long-Short Term Memory (LSTM) models, complement traditional methods. Our results show that machine learning methods fit reasonably well with the data but do not outperform systematically classical methods such as Autoregressive Integrated Moving Average (ARIMA) or the naïve models in terms of out of sample forecasts. However, averaging the forecasts from the two type of models provide better results compared to either method. Compared to the ARIMA and the LSTM, the Root Mean Squared Error (RMSE) of the average forecast was 0.21 and 21.49 percent lower, respectively, for cotton. For oil, the forecast averaging does not provide improvements in terms of RMSE. We suggest using a forecast averaging method and extending our analysis to a wide range of commodity prices.
format Artículo preliminar
id CGSpace143474
institution CGIAR Consortium
language Inglés
publishDate 2021
publishDateRange 2021
publishDateSort 2021
publisher International Food Policy Research Institute
publisherStr International Food Policy Research Institute
record_format dspace
spelling CGSpace1434742025-12-02T21:02:52Z Forecasting commodity prices using long-short-term memory neural networks Ly, Racine Traoré, Fousseini Dia, Khadim models forecasting neural networks commodities cotton machine learning networks oils prices This paper applies a recurrent neural network (RNN) method to forecast cotton and oil prices. We show how these new tools from machine learning, particularly Long-Short Term Memory (LSTM) models, complement traditional methods. Our results show that machine learning methods fit reasonably well with the data but do not outperform systematically classical methods such as Autoregressive Integrated Moving Average (ARIMA) or the naïve models in terms of out of sample forecasts. However, averaging the forecasts from the two type of models provide better results compared to either method. Compared to the ARIMA and the LSTM, the Root Mean Squared Error (RMSE) of the average forecast was 0.21 and 21.49 percent lower, respectively, for cotton. For oil, the forecast averaging does not provide improvements in terms of RMSE. We suggest using a forecast averaging method and extending our analysis to a wide range of commodity prices. 2021-02-01 2024-05-22T12:14:23Z 2024-05-22T12:14:23Z Working Paper https://hdl.handle.net/10568/143474 en Open Access application/pdf International Food Policy Research Institute Ly, Racine; Traore, Fousseini; and Dia, Khadim. 2021. Forecasting commodity prices using long-short-term memory neural networks. IFPRI Discussion Paper 2000. Washington, DC: International Food Policy Research Institute. https://doi.org/10.2499/p15738coll2.134265.
spellingShingle models
forecasting
neural networks
commodities
cotton
machine learning
networks
oils
prices
Ly, Racine
Traoré, Fousseini
Dia, Khadim
Forecasting commodity prices using long-short-term memory neural networks
title Forecasting commodity prices using long-short-term memory neural networks
title_full Forecasting commodity prices using long-short-term memory neural networks
title_fullStr Forecasting commodity prices using long-short-term memory neural networks
title_full_unstemmed Forecasting commodity prices using long-short-term memory neural networks
title_short Forecasting commodity prices using long-short-term memory neural networks
title_sort forecasting commodity prices using long short term memory neural networks
topic models
forecasting
neural networks
commodities
cotton
machine learning
networks
oils
prices
url https://hdl.handle.net/10568/143474
work_keys_str_mv AT lyracine forecastingcommoditypricesusinglongshorttermmemoryneuralnetworks
AT traorefousseini forecastingcommoditypricesusinglongshorttermmemoryneuralnetworks
AT diakhadim forecastingcommoditypricesusinglongshorttermmemoryneuralnetworks