Seasonal Analysis for Globally Prices of Some Agricultural Commodity for the Purpose for Forecasting Using the SARIMA Model

Abstract

Prices are one of those main axes that have an impact on the movement of local and world markets, Therefore, the research aims to forecast the monthly future prices of oranges and soybeans in the world market, To achieve the aims of the research, it relied on both descriptive statistical methods and quantitative methods represented by the univariate times series models of the standard prediction model (SARIMA), the most important results that were reached with regard to the world price that the SARIMA (1,1,1)(2,1,0) model is the best in predicting orange prices, where it passed the tests of estimation and judgment on the predictive quality of the model, including (ADF), (AIC), (U) and (MAPE), where the expected world price of oranges during the period from 8/2022 to 9/2023 ranged between a minimum of $0.739 per kilogram during the month of April 2023 and a maximum of about $0.784 per kilogram during the month of September 2022, with an error of 13.4%, With regard to the global price of soybeans, the estimates of the SARIMA model (2,1,2)(0,1,2) are the best in the future forecast, Where the expected world price of soybeans during the period from 8/2022 to 9/2023 ranged between 574.9 dollars per ton as a minimum during the month of May of 2023 and a maximum of about 656.1 dollars per ton as a maximum during the month of August of 2022, with an error of 16.4%, therefore.

Keywords

• stationary
• univariate time-series models
• sarima model
• augmented dickey fuller test
• theil coefficient
• forecast