Application of long short term memory to forecast tomato prices

Abstract

Fluctuations in agricultural commodity prices affect the supply and demand of agricultural products and have a significant impact on consumers so accurate prediction of agricultural commodity prices would facilitate the reduction of risk caused by price fluctuations. This paper applies an Artificial Neural Network (ANN) method to forecast tomato prices. We showed how this new tool from machine learning, particularly Long-Short Term Memory (LSTM) models. Here we used LSTM model to implement price prediction as it is a best model for time series data. The research question investigated in this article is that whether and how the newly developed deep learning based algorithms for forecasting time series data, such as “Long Short-Term Memory (LSTM)”, are superior to the traditional algorithms. The empirical studies conducted and reported in this article showed that deep learning-based algorithms such as LSTM outperform traditional-based algorithms such as ARIMA and Holt-Winter Seasonal Method. From the result, LSTM is discovered to be the most accurate and efficient in handling increasing amounts of complex data. To evaluate the LSTM, weekly field prices of tomato for 10 years (20011-202) were used. The results showed that, compared to the ARIMA and the Holt-Winter Seasonal Method, the Root Mean Squared Error (RMSE), Mean Absolute Percentage Error (MAPE) and Mean Absolute Error (MBE) of LSTM showed lesser error value i.e., 267.66, 0.15 and 176.40 respectively. Thus, it provides a new idea for agricultural commodity price forecasting and has the potential to stabilize the supply and demand of agricultural products, thus help to understand the price fluctuation and also it provides an insight for the government in order to take necessary decisions for managing risk.