Statistical Modeling and Forecasting of Area, Production and Yield of Mango

Abstract

The goal of the current study was to forecast and analyze the trends in the area, production and yield of mango crops planted in Andhra Pradesh between 1991-1992 and 2017-2018. Different non-linear and ARIMA models were used to forecast and analyse area, production, and yield trends. The cubic model for production and yield of the mango crop farmed in Andhra Pradesh and the Power model for area trends both fit the data well. For forecasting mango area, production and yield, the best models were ARIMA (2,0,2), Cubic and ARIMA (2,2,1) based on model selection parameters such as adjusted R2 , RMSE, and MAE. The findings showed that the yield, production and area of the growing mango crop in Andhra Pradesh would likely exhibit an upward trend in the future years. The projected figures for 2022-2023 were roughly 641 thousand hectares of mango area, 6565 thousand metric tonnes of production and 45.4 mt/ha of mango crop yield. A non-significant positive correlation was found between the area and mango yield.. KEYWORDS :Growth rates, Non-linear, ARIMA models, Ljung-Box Q test, 2 R , RMSE, MAE and Correlation.