Hybrid SARIMA-Facebook Prophet model for prediction and forecasting of the staple food prices in Tanzania

Abstract

The country's stable development requires sufficient food production and reserves for its people. However, the global climatic change and other factors, including political instability, resulted in the shortage and volatility of staple food prices. Thus, the best prediction and forecasting models of staple food prices become essential. The advancement of technology gave room for the development of hybrid models. Therefore, in this paper, we develop the SARIMA-Facebook (FB) Prophet model to predict and forecast staple food prices in Tanzania. Based on the Root Mean Squared Error (RMSE) metrics, the proposed hybrid model out performs the SARI MA and other hybrid models in predicting staple food prices. Moreover, the forecasts for the next two years have shown increasing trends in staple food prices. Since most neighbouring countries depend on Tanzania to complement the food crisis in their countries, initiatives should be taken into account to stabilize prices. Finally, the proposed hybrid SARIMA-FB Prophet may be applied to predict and forecast other prices to see the stability of its accuracy. The construction of the theoretical hybrid SARIMA-FB Prophet model may be done in future studies.